Anticipated
acquisition by
Amazon of a minority
shareholding and
certain rights in
Deliveroo
Final report
4 August 2020
© Crown copyright 2020
You may reuse this information (not including logos) free of charge in any format or
medium, under the terms of the Open Government Licence.
To view this licence, visit www.nationalarchives.gov.uk/doc/open-government-
licence/ or write to the Information Policy Team, The National Archives, Kew, London
Website: www.gov.uk/cma
Members of the Competition and Markets Authority
who conducted this inquiry
Stuart McIntosh (Chair of the Group)
Humphrey Battcock
Paul Hughes
Claire Whyley
Chief Executive of the Competition and Markets Authority
Andrea Coscelli
The Competition and Markets Authority has excluded from this published version
of the final report information which the inquiry group considers should be
excluded having regard to the three considerations set out in section 244 of the
Enterprise Act 2002 (specified information: considerations relevant to disclosure).
The omissions are indicated by []. Some numbers have been replaced by a
range. These are shown in square brackets. Non-sensitive wording is also
indicated in square brackets.
1
Contents
Page
Summary .................................................................................................................... 4
Findings .................................................................................................................... 28
1. The reference ..................................................................................................... 28
2. Coronavirus (COVID-19) outbreak ...................................................................... 29
Effect of the Coronavirus (COVID-19) outbreak on the UK economy ................. 29
Effect of the Coronavirus (COVID-19) outbreak on the restaurant sector ........... 30
Effect of Coronavirus (COVID-19) in March 2020 .......................................... 30
Effect of Coronavirus (COVID-19) from mid-April 2020 ................................. 31
Effect of the Coronavirus (COVID-19) outbreak on grocery sector ..................... 32
Longer-term effect of Coronavirus (COVID-19) .................................................. 33
3. The Parties, the Transaction and the rationale ................................................... 33
The Parties ......................................................................................................... 33
Amazon ......................................................................................................... 33
Deliveroo ....................................................................................................... 37
The Transaction .................................................................................................. 40
The rationale ....................................................................................................... 41
Amazon’s rationale ........................................................................................ 41
Deliveroo’s rationale ...................................................................................... 43
4. Jurisdiction .......................................................................................................... 45
Enterprises ceasing to be distinct ....................................................................... 46
Arrangements are in progress or in contemplation ........................................ 46
Enterprises .................................................................................................... 46
Ceasing to be distinct .................................................................................... 47
Influence through Amazon’s shareholding in Deliveroo ................................. 49
Influence through Amazon’s right to representation on Deliveroo’s board ..... 57
Other sources of influence ............................................................................. 60
Overall assessment of material influence ........................................................... 62
Conclusion on relevant merger situation ............................................................. 64
5. Market definition .................................................................................................. 64
Approach ............................................................................................................ 64
Online restaurant platforms ................................................................................. 65
Introduction and overview .............................................................................. 65
Restaurant platform models........................................................................... 66
Product market definition ............................................................................... 67
Geographic market definition ......................................................................... 85
Online convenience groceries ............................................................................. 89
Introduction and overview .............................................................................. 89
Product market definition ............................................................................... 91
Geographic market definition ....................................................................... 108
6. Counterfactual................................................................................................... 111
Legal framework ............................................................................................... 111
Counterfactual for online convenience groceries ......................................... 112
Deliveroo counterfactual ................................................................................... 113
Overview of exiting firm counterfactual ........................................................ 113
Summary of April Provisional Findings analysis .......................................... 113
Initial impact of Coronavirus (COVID-19) on Deliveroo ............................... 114
Impact of Coronavirus (COVID-19) on the April Provisional Findings
counterfactual ............................................................................................. 115
2
Developments since the April Provisional Findings ..................................... 117
Impact of developments on Deliveroo ......................................................... 118
Parties’ submissions .................................................................................... 120
Third party submissions ............................................................................... 122
CMA assessment of the Deliveroo counterfactual ....................................... 123
Limb 1: Would Deliveroo have exited absent the Transaction? ................... 123
Conclusion on the application of ‘Limb 1’ of the exiting firm analysis .......... 129
Limb 2: Is there a substantially less anti-competitive investor or purchaser for
Deliveroo’s assets? .................................................................................... 129
Amazon counterfactual ..................................................................................... 131
Parties’ submissions .................................................................................... 133
CMA assessment: Amazon’s incentives and intention to re-enter ............... 134
CMA assessment: Amazon’s ability to re-enter ........................................... 157
Conclusion on re-entry by Amazon counterfactual ...................................... 168
7. Supply of online restaurants platforms in the UK .............................................. 168
Framework ........................................................................................................ 169
Parties’ submissions ......................................................................................... 170
Would Amazon re-enter? .................................................................................. 171
Theories of harm ............................................................................................... 172
Impact of the Transaction on whether Amazon would re-enter ................... 175
Impact of the Transaction on the nature of Amazon’s entry ........................ 183
Impact of Amazon’s re-entry ............................................................................. 187
Conclusion on supply of online restaurant platforms in the UK ......................... 189
8. Supply of online convenience grocery services in the UK ................................. 190
Introduction ....................................................................................................... 190
Background ....................................................................................................... 191
The groceries industry ................................................................................. 191
How OCG services are supplied .................................................................. 194
Closeness of competition .................................................................................. 198
The Parties’ current OCG businesses ......................................................... 198
Substitutability of the Parties’ OCG propositions ......................................... 203
The Parties’ ambitions in groceries .............................................................. 212
The Parties’ plans to supply OCG ............................................................... 214
Competition in OCG .......................................................................................... 216
Online grocery providers ............................................................................. 218
Online restaurant delivery marketplaces ..................................................... 219
Convenience store providers ....................................................................... 221
Traditional grocery stores ............................................................................ 222
Expansion and development of competitive offers ...................................... 224
Future competitive constraints on Amazon and Deliveroo in OCG .............. 236
Theories of harm ............................................................................................... 241
Amazon discouraging Deliveroo from competing against Amazon in OCG . 245
Amazon competing less aggressively in OCG services ............................... 247
Amazon relying on Deliveroo for its presence in OCG ................................ 249
9. Additional theories of harm ............................................................................... 252
Bundling of Deliveroo in Amazon Prime ............................................................ 252
Third party views ......................................................................................... 253
Our assessment .......................................................................................... 254
Preventing third party access to Deliveroo’s delivery network .......................... 259
Third party views ......................................................................................... 259
Our assessment .......................................................................................... 261
4
Summary
Overview
1. The Competition and Markets Authority (CMA) has found that the anticipated
acquisition by Amazon.com NV Investment Holdings LLC, a wholly-owned
subsidiary of Amazon.com, Inc (Amazon) of certain rights and a 16% minority
shareholding in Roofoods Ltd (Deliveroo) (the Transaction) has not resulted,
and may not be expected to result, in a substantial lessening of competition
(SLC) within a market or markets in the United Kingdom (UK) for goods and
services.
2. On 27 December 2019, the CMA referred the Transaction for further
investigation and report by a group of CMA panel members (the Inquiry
Group) following a phase 1 review. The terms of reference said that the
Inquiry Group must decide:
(a) whether arrangements are in progress or in contemplation which, if
carried into effect, will result in the creation of a relevant merger situation;
and
(b) if so, whether the creation of that situation may be expected to result in an
SLC within any market or markets in the UK for goods or services.
1
3. In its phase 1 investigation, based on the information available at that time,
the CMA found that it is or may be the case that the Transaction may be
expected to result in an SLC in (a) the supply of online restaurant platforms in
the UK; and (b) the supply of online convenience grocery platforms in the UK.
We therefore focussed on these areas in our inquiry.
4. We also considered whether we should further investigate any of the theories
of harm that were dismissed following the phase 1 investigation, in particular,
the possibility that the Parties
2
could choose to bundle their respective
services, potentially foreclosing other companies from competing with them.
We found no additional evidence suggesting that these theories of harm
should be re-investigated at phase 2.
1
See Terms of reference, 27 December 2019.
2
Throughout this document we refer to Amazon and Deliveroo collectively as ‘the Parties’.
5
Background
5. We published two sets of provisional findings in the course of the phase 2
investigation of the Transaction (the April Provisional Findings and Revised
Provisional Findings).
3
We received detailed and considered responses to
those provisional findings from many stakeholders, and those responses have
been helpful to the Inquiry Group in considering the issues presented by the
Transaction and reaching our final decision.
6. In late-March 2020, Deliveroo advised the CMA that, if the company’s
directors did not have a reasonable expectation of receiving additional funds
before Deliveroo ran out of cash in early Q3 2020, they would shortly be
required to initiate insolvency proceedings. The evidence gathered by the
CMA at the time, in particular Deliveroo’s financial data provided in early-April
2020 (ie actual data up to the beginning of April and forecast data based on
the situation at that time), showed that Coronavirus (COVID-19) was having a
severe impact on Deliveroo’s business. We understood that the situation was
urgent and, therefore, sought to publish provisional findings as quickly as
possible.
4
Given the conclusion at that time, that Deliveroo was likely to exit
the market unless it received the additional funding available through the
Transaction, and that the situation would have been the same in the
counterfactual, it was not necessary to set out our full analysis of the effect of
the Transaction on competition in the supply of goods and services in the UK.
7. Beginning in April 2020, and accelerating through that month, market
conditions and Deliveroo’s financial situation changed materially leading us to
conclude Deliveroo was no longer likely to exit the market as a result of the
Coronavirus (COVID-19) pandemic. In light of this finding, we published our
full provisional findings setting out our analysis of the effects of the
Transaction on competition against a counterfactual where Deliveroo remains
in the market (the Revised Provisional Findings).
5
It was our view that the
Transaction would not be expected to result in an SLC in either the market for
online restaurant platforms or the market for online convenience groceries
(OCG) in the UK.
8. Following a review of the responses to the Revised Provisional Findings, and
further information provided by the Parties, we have reached our final
decision, which is set out in detail below.
3
See April Provisional Findings, 16 April 2020 and Revised Provisional Findings, 22 June 2020.
4
See April Provisional Findings, 16 April 2020.
5
See Revised Provisional Findings, 22 June 2020.
6
The Parties
9. Founded in 1994 by current CEO Jeff Bezos, Amazon has grown rapidly to
become one of the world’s largest companies with a market capitalisation of
c.$1.52 trillion.
6
In 2019, Amazon had over $280 billion
7
in net sales (up 20%
year on year), over $14 billion in operating income
8
and approximately
800,000 employees.
9
‘Amazon.com, Inc.’, the parent company of the group, is
publicly listed on the US Nasdaq Global Select Market (NASDAQ) and
incorporated in Delaware in the United States (US).
10. The UK is Amazon’s third largest market in terms of net sales (behind the US
and Germany) with approximately 6% of net sales generated from the UK
(approximately $17.5 billion in 2019).
10
Amazon’s activities include operating
as an online retailer and third-party marketplace provider, a delivery and
logistics network operator, a host of cloud server space, a book publisher, a
producer and online broadcaster of television and films and a manufacturer of
electrical devices such as Amazon Kindle e-readers and Echo devices.
11. Deliveroo is a UK-based company founded in 2013 whose main activity is
restaurant delivery. In addition to restaurant delivery, the company operates
cloud kitchens through Deliveroo Editions, delivers convenience groceries
through a partnership with Co-op and certain other retailers, and provides
wholesale and data services to restaurants. Deliveroo’s primary market is the
UK, where it operates in over 200 towns and cities, working with over 30,000
partner restaurants and 25,000 riders and employs over 1,000 people directly.
The Transaction
12. The Transaction comprises [] which Amazon entered into with Deliveroo on
16 May 2019:
11
(a) [], which gives Amazon certain rights over Deliveroo, including board
representation, a []% shareholding ([]); and
(b) [], which envisages the acquisition by Amazon of a []% shareholding
(and certain other rights) in Deliveroo, [].
12
6
As at 28 July 2020.
7
Amazon.com, Inc Q4 2019 Financial Results Conference Call Slides, slide 8. In Q1 2020, Amazon had over
$75 billion in net sales (Amazon.com, Inc Q1 2020 Financial Results Conference Call Slides, slide 7).
8
Amazon.com, Inc Q4 2019 Financial Results Conference Call Slides, slide 10. In Q1 2020, Amazon had around
$4 billion in operating income (Amazon.com, Inc Q1 2020 Financial Results Conference Call Slides, slide 9).
9
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, page 4.
10
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, page 68.
11
[].
12
The German authorities cleared this Transaction on 11 July 2019.
7
Jurisdiction
13. We have found that the Transaction has resulted in the creation of a relevant
merger situation within the meaning of the Act on the basis that, as a result of
the Transaction, Amazon will have the ability to exercise material influence
over Deliveroo and Deliveroo has revenues in the UK in excess of £70 million.
14. An assessment of material influence requires a case by case analysis of the
overall relationship between the acquirer and the target. When assessing
material influence, the CMA will consider the significance of the rights
conferred by formal agreements and structures, as well as the commercial
realities of the relationship between an acquirer and the target.
15. Our view is that the following factors, taken together, demonstrate that
Amazon will acquire material influence over Deliveroo:
(a) the size of Amazon’s investment (in both absolute terms and relative to
other shareholders) and its associated rights;
(b) a strong body of evidence that Deliveroo’s management, its other
shareholders and its commercial/operations teams perceive that Amazon:
(i) has a special status as a significant ‘strategic’ investor, with various
additional rights, and is a credible potential future acquirer of
Deliveroo (or source of funding);
(ii) has commercial and operational expertise from running an online
business in directly relevant sectors to Deliveroo; and
(iii) is a current and potential future strategic/commercial partner of
Deliveroo.
16. Together, we consider these factors will mean Amazon’s views are likely to be
given material weight by both management and other shareholders (and their
appointed directors), such that we conclude that the Transaction will give
Amazon the ability materially to influence decisions related to the policy of
Deliveroo, including the management of its business, the strategic direction of
the company and/or its ability to define and achieve its commercial objectives.
Market definition
17. Market definition provides a framework for assessing the competitive effects
of a merger for a relevant product and geographic market. We have
considered the impact of the Transaction on the supply of:
8
(a) online restaurant platforms in the UK; and
(b) online convenience grocery delivery in the UK, that is, groceries that are
delivered within a short period of time after ordering.
Counterfactual
18. The counterfactual is an analytical tool used to help answer the question of
whether a merger has or may be expected to result in an SLC. It does this by
providing the basis for a comparison of the competitive situation on the market
with the merger against the most likely future competitive situation on the
market absent the merger.
13
The latter is the counterfactual.
19. Where there is more than one possible alternative scenario, during a phase 2
review the CMA will select the counterfactual it considers would be the most
likely scenario to have arisen absent the merger.
20. Against this framework and in light of the Parties’ submissions and the
evidence we observed, we considered the likely future situations of both
Amazon and Deliveroo in the absence of the Transaction. Our focus was on
two key questions: (a) whether Deliveroo would exit the market in the absence
of the Transaction, as Deliveroo argued prior to the April Provisional Findings;
and (b) whether Amazon would re-enter the market for online restaurant
platforms. We have also considered the relevant counterfactual in the market
for OCG.
Exiting Firm Counterfactual
21. When considering an ‘exiting firm’ scenario, the CMA examines whether the
firm would have left the market and whether the transaction at issue is the
best available outcome for consumers.
14
The CMA applies a three-limb test
when making this assessment, considering:
(a) Whether the firm would have exited (through failure or otherwise) absent
the transaction.
(b) Whether there would have been an alternative purchaser for the firm or its
assets.
(c) What the impact of exit would be, and how this would compare to the
impact of the transaction.
13
Merger Assessment Guidelines (CC2 Revised) (MAGs), paragraphs 4.3.1 and 4.3.6.
14
See April Provisional Findings, 16 April 2020.
9
22. As a loss-making business, Deliveroo has historically been reliant on external
funding in order to continue trading. We do not, however, consider that
Deliveroo would be a failing firm in the counterfactual on this basis. This is
consistent with the Parties’ submission in the Merger Notice that, absent the
Transaction, Deliveroo ‘would have continued to compete as it currently does,
including by seeking suitable investment to drive its expansion and
innovation’.
23. In the April Provisional Findings, we provisionally concluded that Deliveroo
would have been likely to exit the market as a result of the Coronavirus
(COVID-19) crisis without additional funds and that, as a result of Deliveroo’s
urgent need for alternative investment and the effect of the crisis on funding
markets, Deliveroo was unlikely to be able to raise the funding required in the
time it had available. We also provisionally concluded that Deliveroo exiting
the market would have had a greater negative effect on competition and
consumers than any effect from allowing the Transaction to proceed.
15
24. Following publication of the April Provisional Findings, we found evidence that
market conditions and Deliveroo’s financial position had changed. As a result,
we no longer considered that Deliveroo should be considered an exiting firm.
Limb 1: whether Deliveroo would exit absent the Transaction
25. Updated financial information provided by Deliveroo in response to requests
for information from the CMA after publication of the April Provisional Findings
shows that Deliveroo’s actual performance in April, May and June was
significantly better than Deliveroo had forecast at the beginning of April. The
recovery appears to have begun slowly in early-April 2020 and accelerated
later in the month.
26. Although the Coronavirus (COVID-19) crisis initially had a severe impact on
Deliveroo’s order volumes, restaurant availability and, in turn, cash flow, this
impact was short-lived. Deliveroo’s actual performance in April, May and June
2020 was significantly better than had been forecast and indeed Deliveroo
was both profitable and had a positive cashflow in May and June as opposed
to the negative cashflow it had forecast.
16
This improved cashflow provides
Deliveroo with more time to seek additional funding and to rationalise its
business operations, potentially extending this cash runway further. The CMA
considers that the improvement in Deliveroo’s performance is primarily a
15
See April Provisional Findings, 16 April 2020.
16
Deliveroo was profitable excluding [] which do not affect net assets.
10
result of changing market conditions, and would also have occurred in the
counterfactual.
27. We do not agree with Deliveroo’s submission that the evidence that
Coronavirus (COVID-19) has had a less severe impact on its business than
originally forecast is irrelevant to our consideration of the counterfactual. In
the absence of evidence of a severe and ongoing financial decline in
Deliveroo’s actual business as a result of the impact of Coronavirus (COVID-
19), we have no basis for finding that such a decline would have occurred in
the counterfactual (ie had Deliveroo accepted an alternative investment to
Amazon).
28. Based on the evidence we have gathered that Coronavirus (COVID-19) has
had a more limited impact than expected on Deliveroo’s business, and
notwithstanding the company’s historic reliance on external funding, our view
is that the most likely counterfactual is the one proposed by the Parties in the
Merger Notice: that, absent the Transaction, Deliveroo ‘would have continued
to compete as it currently does, including by seeking suitable investment to
drive its expansion and innovation’. On that basis, we no longer consider that
Deliveroo would be likely to exit the market absent the Transaction.
Limb 2: whether there would be alternative investors in Deliveroo
29. On the basis of the evidence we had at the time of the April Provisional
Findings, we believed that Deliveroo was facing a financial ‘cliff-edge’ as a
result of the significant impact of Coronavirus (COVID 19) with no evidence of
a likely improvement in the near future. That situation would have required
funds to be raised in a very short period of time, during the period that
investment sentiment was weakest.
30. While we have received evidence that it would have been challenging for
Deliveroo to raise funds in March 2020, our view is that in the relevant
counterfactual Deliveroo would likely have taken action to ensure that it was
both more attractive to potential investors and to seek to avoid a ‘cliff edge’ in
which it would have needed to secure funding very urgently, for example by
raising funds prior to March 2020. In addition, for the reasons set out above,
the urgent funding need due to Coronavirus (COVID-19) envisaged in the
April Provisional Findings no longer appears likely to have arisen.
31. We consider that the evidence that the impact of Coronavirus (COVID-19) on
both Deliveroo and the funding markets has been more muted than expected
at the date of the April Provisional Findings means that the most likely
counterfactual absent the Transaction is that Deliveroo would have continued
11
to compete in the market, and to raise funds to do so, in the same way as it
anticipated prior to the crisis.
Restaurants counterfactual
32. We have examined whether the most likely scenario, absent the Transaction,
is that Amazon would choose to re-enter the supply of online restaurant
platforms in the UK. Amazon previously had a limited presence in this market
and exited in 2018. In order to assess the likelihood of Amazon re-entering
the supply of online restaurant platforms in the UK, we have assessed
Amazon’s incentives and intention to re-enter, primarily by examining
evidence from Amazon’s internal documents. We then assessed Amazon’s
ability to re-enter, again by examining Amazon’s internal documents and
looking at evidence from third parties.
33. When considering whether a business has the incentive, intention and ability
to enter a market, it is not necessary to identify a single specific route by
which that business would be more likely than not to enter the market. Nor do
we consider it necessary to have internal documentary evidence setting out
an explicit, concrete intention to enter within a defined timeframe. Rather, we
must undertake an in-the-round assessment that reflects all of the available
evidence with respect to a party’s intention, incentive and ability to enter.
34. We have taken into account a range of evidence, including evidence of the
way that Amazon operates its business in practice (and in particular its so-
called ‘test and learn’ approach to innovating and expanding its product
offerings). We have carefully considered evidence relating to the broader
commercial strategy of the Amazon business, and the perceived potential
importance of the online restaurant platform market within this strategy. Our
assessment was also informed by Amazon’s current strategies for Prime, and
its assessment of the role that food offerings play within these strategies,
evidence relating to Amazon’s previous attempt at supplying a restaurant
platform (Amazon Restaurants), and evidence relating to its interest and
subsequent investment in Deliveroo.
Amazon Prime
35. Promoting and growing Prime is very important to Amazon, and it has
successfully continued to grow Prime membership globally in recent years.
After the US, the UK is Amazon’s [] largest territory in terms of Prime
subscription revenue and Prime household penetration. The number of
Amazon Prime customers in the UK has grown rapidly in recent years.
Amazon has invested heavily in its Prime proposition in the UK, and part of its
[].
12
Offering restaurant delivery as part of Amazon Prime
36. Food, in general, appears to be an area which Amazon sees as high-use and
of value to its Prime customers,
17
and []. Restaurant delivery is seen as a
useful benefit for Prime. First, this was shown by Amazon’s expectations of
Amazon Restaurants. Second, the attraction and benefit of restaurant delivery
[] for discussion with Jeff Bezos (CEO) and other members of Amazon’s
senior leadership team.
37. The UK is one of the largest markets for restaurant delivery in the world and is
the largest in Europe. Amazon would have a significant advantage over other
operators seeking to enter the UK as it can benefit from its existing
relationships with millions of customers in the UK, including engaged Prime
customers. We also consider that Amazon may have a different time horizon
for profitability compared to other potential entrants and the financial
resources to support this.
38. There is evidence that Amazon sees offering fast delivery of a range of
products as a way to enhance the value of Prime, which the evidence shows
is important to Amazon’s overall global corporate strategy. Amazon has
considered restaurant delivery on a global level in the context of its desire to
attract and retain the customers of its Prime subscription service and to be
known for fast delivery. Amazon has a global strategy, as explained below, of
expanding its grocery offering and increasing the speed of delivery.
Restaurant delivery through Amazon Restaurants was expected to play an
important role in this. The benefits Amazon has identified from restaurant
delivery in internal documents include: [].
Amazon’s food strategy
39. F3 is an Amazon business area which used to also include Amazon
restaurants. F3 operates internationally and is focused on developing
Amazon’s online grocery offering including through the roll-out of its ultra-fast
grocery (UFG) plan.
40. Several internal emails from within F3 refer to the importance of offering
restaurant delivery as part of Amazon’s food strategy, and in particular using
the assets or expertise of Deliveroo in other parts of Amazon’s food
businesses. In addition, emails from the corporate development team show
17
For instance, Amazon announced in July 2020 that its Amazon Fresh grocery delivery service would now be
available for free for Prime members. See: Amazon Fresh Homepage.
13
consideration of wider benefits to Amazon from acquiring Deliveroo including
[].
41. Evidence shows the current priority for F3 is to expand Amazon’s position in
online groceries globally, with a strong focus on [], and it is investing heavily
in its delivery and distribution capabilities in the US and UK. Alongside this,
Amazon’s interest in restaurant delivery has continued for the past five years,
is international, and extends across the F3 team, as well as to senior
executives at the highest level within Amazon, including Jeff Bezos. There is
no evidence to suggest that Amazon is no longer interested in restaurant
delivery or that it no longer expects it to be an important area providing
benefits such as differentiation in its offering, flywheel
18
effects for Prime, and
enhanced logistical capabilities.
42. While Amazon is currently focused on its groceries business, capabilities it is
developing in this context would also support restaurant food delivery. For
example, as part of its US grocery offering, [], and has further plans to
increase the speed of its grocery offering. Delivery speed and [] were
among the main challenges Amazon faced when operating Amazon
Restaurants.
43. We have considered whether the effect of the Coronavirus (COVID-19)
pandemic would change this assessment. Although the longer term effects of
Coronavirus (COVID-19) are not clear, based on the data available at this
time, we have no evidence to suggest that the pandemic will make the UK a
less attractive country for Amazon to operate an online restaurant platform, or
that Amazon’s incentive or ability to enter this segment will be significantly
negatively affected. Therefore, we do not consider that the effect of the
pandemic changes our assessment of Amazon’s food strategy or its
incentives to re-enter restaurant delivery.
Amazon Restaurants
44. Amazon Restaurants was trialled in Seattle, and then launched more widely in
2015, when Amazon started building its own delivery service for restaurants
and added Amazon Restaurants to its Prime Now offering. Amazon
Restaurants was launched in the UK in 2016. Amazon Restaurants closed in
the UK in 2018 and in the US in 2019.
18
The term ‘flywheel’ is used at Amazon to describe something similar to a virtuous cycle, which powers the
business. For example, lower prices lead to more customer visits, more customer visits increase the volume of
sales, and that results in more commission-paying third-party sellers to the site. [].
14
45. Evidence of Amazon’s expectations for its restaurant delivery business shows
that Amazon considered the online restaurant platform business to have
strategic value for the wider Amazon business. Amazon’s expectations
included that restaurant delivery would be an important part of a wider food
offering for consumers, it would be an [], and that the technology used for
the fast delivery of restaurant food would benefit the wider Amazon business
as it could be rolled out to other areas, including grocery and retail goods, in
the UK, US and elsewhere.
46. The evidence shows that Amazon appears to have made some operational
errors and strategic misjudgements in the way it implemented Amazon
Restaurants in the UK. []. Although it took some remedial action, [], it
was not sufficient to turn the struggling business around to the satisfaction of
Amazon’s senior executives. Amazon Restaurants also struggled in the US
and Amazon did not appear to be willing to invest significantly more in order to
achieve scale in the market, given the preference ultimately expressed by Jeff
Bezos [].
47. Internal documents prepared by Amazon when it closed the restaurants
business indicate the potential for Amazon to re-enter this area [] gained
from Amazon Restaurants across different countries. Evidence from internal
documents also shows [].
48. We do not consider that Amazon’s decision to exit the UK means that it was
no longer interested in this market – indeed, just months after the closure of
the UK business Amazon began due diligence on a potential transaction with
Deliveroo, suggesting a continuing interest in the UK restaurant platform
market.
49. Furthermore, the decision to close the Amazon Restaurants business in the
US and the decision to invest in Deliveroo were taken almost simultaneously,
and both decisions were taken by the CEO, Jeff Bezos. []. Even though
[], evidence indicates that it sees restaurant delivery as ‘strategically
aligned’ with its business overall.
50. Although Amazon has submitted that the decisions to invest in Deliveroo and
to close Amazon Restaurants were taken separately, the CMA considers that
there is evidence suggesting these decisions were linked. Given the
appearance that these two decisions were linked, the CMA has repeatedly
provided Amazon with opportunities to clarify the rationale for these decisions,
but Amazon failed to do so. We consider that that the evidence indicates that
the decision to close Amazon Restaurants US was influenced by Amazon
having an alternative route to enter restaurant delivery internationally.
15
Conclusion on Amazon’s incentives and intention
51. We believe that Amazon has a strong and continued interest in online
restaurant platforms and an incentive to offer this service to Prime customers
in order to differentiate its offering, realise flywheel benefits, and develop
useful logistical capabilities that would be deployed elsewhere in its business.
Evidence shows that the UK is a large and growing market for online
restaurant platforms, and the UK is an important and attractive market to
Amazon as a result of high levels of Prime membership. Since the closure of
Amazon Restaurants, we have observed in Amazon’s internal documents
continued global interest in online restaurant food delivery.
Amazon’s ability to re-enter
52. There are a number of possible routes that Amazon could take to re-enter the
supply of online restaurant platforms. We believe there are three main ones:
(i) building its own offering; (ii) acquiring or investing in an existing restaurant
platform or adjacent business; and (iii) partnering with an existing restaurant
platform or adjacent business. For the purposes of establishing the
counterfactual in this case, we need only consider the viability of these
options with a view to determining whether the most likely scenario is that
Amazon would enter via one or other of these routes. We do not need to
determine which of these routes would be most likely.
53. As explained above, Amazon has rolled out some of the technology that was
used in Amazon Restaurants to support its F3 business, and has continued to
develop this technology. Amazon is also using technology from Amazon
Restaurants in other areas such as the Amazon India restaurant delivery
business. The continued use of this technology indicates that Amazon
Restaurants allowed Amazon to learn and develop technology that is useful in
operating this type of business, and that the technology is transferrable. Third
parties active in online restaurant delivery have told us that the technology
powering their platform and delivery network is generally transferable between
countries.
54. Amazon could also acquire an overseas online restaurant platform. As
described above, we believe that the evidence shows Amazon has an
incentive to offer this service in the UK based on the attractiveness of the UK
restaurant delivery market and Amazon’s broader presence in the UK. We
believe that if Amazon looked to enter through acquisition, it could offer this
service in the UK in the short to medium-term (ie within five years).
55. A further option would be for Amazon to re-enter the supply of online
restaurant platforms through contractual arrangements with a business that
16
either offers an online restaurant platform already, or that would enable
Amazon to offer an online restaurant platform by addressing one or more of
the barriers to entry that Amazon faces. Evidence shows that companies are
interested in working with Amazon in this market.
56. We consider that the benefit to Amazon of acquiring, investing in, or
partnering with a third party is that one or other of these approaches would
allow it to more easily overcome the barriers to entry in the supply of online
restaurant platforms in the UK. An internal email shows [].
57. In pursuing any route to entry, Amazon would need to attract customers,
restaurants and couriers to its platform. Amazon has access to [] Prime
subscription customers in the UK, a number which has [] compared to when
it launched Amazon Restaurants in the UK. As this is a subscription service
offering a wide range of benefits to subscribers, these customers are likely to
regularly engage with Amazon products. This gives Amazon multiple avenues
for engaging with customers and marketing a new service to them, thus
saving significant amounts on marketing and advertising compared to other
players.
58. Overall, there are multiple possible routes to entry for Amazon and there is
evidence of interest in alternative providers as targets or partners. There exist
a number of potential partners and/or targets, including non-UK restaurant
platforms as well as UK-based logistics specialists, that could help Amazon
overcome the barriers to entry in supplying a restaurant platform in the UK,
including those that hampered its previous attempt in this market. Amazon
could use the learning from Amazon Restaurants to avoid repeating the same
strategic mistakes, or could invest in or partner with an alternative provider to
gain additional expertise in this market.
Conclusion on re-entry by Amazon in the counterfactual
59. On the basis of the evidence set out above, we believe that Amazon has a
continued interest in online restaurant platforms and an incentive to offer this
service to Prime customers in order to differentiate and add to its offering and
develop useful logistics capabilities that would be deployed elsewhere in its
business. Although much of the evidence considered relates to international
markets, we consider that the UK is an attractive market for restaurant
delivery and is an important market for Amazon. We consider the evidence
shows there are multiple possible routes for entry for Amazon absent the
Transaction. While entry into the UK market by Amazon may not occur
imminently, we consider that, absent the investment in Deliveroo, the most
likely scenario would involve Amazon choosing to re-enter in the short to
medium term (ie within five years). We therefore conclude that, in the
17
counterfactual, Amazon is likely to re-enter the supply of online restaurant
platforms in the UK.
Groceries counterfactual
60. With respect to the market for OCG, we considered whether the development
and expansion plans of the Parties should be taken into account in the
counterfactual. The online convenience grocery market is nascent, and many
market participants are at an early stage of trialling their offerings, or gradually
rolling them out across geographical areas. Views vary considerably as to
how the market will develop. In our view, future OCG market developments
are not sufficiently foreseeable to include in a counterfactual. We have,
however, taken account of possible changes in the market, including the
expansion plans of the Parties and their competitors, in our competitive
assessment. We have therefore adopted a counterfactual in which the Parties
would have continued to develop and strengthen their OCG propositions
independently of one another absent the Transaction.
Supply of online restaurants platforms in the UK
61. The CMA’s Merger Assessment Guidelines provide a framework for the
assessment of unilateral effects arising from the loss of potential competition
as a result of a merger. The guidance sets out two questions to be addressed
when considering ‘actual potential competition’:
19
(a) would the potential entrant be likely to enter in the absence of the merger;
and
(b) would such entry lead to greater competition?
62. The guidance also states that the CMA will consider whether there are other
potential entrants before reaching a conclusion on the SLC test.
63. In line with the CMA’s guidance on assessing the loss of potential
competition, we considered (a) whether Amazon is likely to enter, and
(b) would this entry lead to greater competition.
64. In order to assess whether Amazon would re-enter the supply of online
restaurant platforms in the UK, we have considered evidence around its
intention, incentives and ability to re-enter. As set out above, we believe that
19
MAGs, paragraph 5.4.15.
18
in the counterfactual Amazon would be likely to re-enter the supply of online
restaurant platforms in the UK in the absence of the Transaction.
65. For the purposes of our competitive assessment, we must compare the
situation arising in the counterfactual to the situation arising as a result of the
Transaction and consider what impact the Transaction has on the potential
future competition between Amazon and Deliveroo. This assessment must
take into account the fact that the Transaction is the acquisition of a 16%
shareholding as opposed to an acquisition of a larger stake, which (together
with other rights) might give rise de facto control or a full merger.
66. If Amazon had acquired 100% of Deliveroo, the acquisition would result in
alignment of incentives between Amazon and Deliveroo and we would not
expect there to be any future competition between them because their
incentives were aligned. We would not expect Amazon still to enter
independently as well as to acquire Deliveroo outright.
67. It is possible that a 16% shareholding could change Amazon’s incentives such
that it would no longer enter independently. It is also possible that Amazon
could enter the market independently while retaining a minority shareholding
in Deliveroo, leading to Amazon having two competing investments in the
same market. The effect of a 16% shareholding could vary depending on the
circumstances of the case, for example the rights that accompany the
shareholding, the strategies of the businesses involved, the nature of the
markets at issue and the constraints from other competitors in those markets.
68. We have considered two scenarios through which harm could arise as a result
of the Transaction:
(a) unilateral effects on the entry decision (meaning the 16% shareholding is
Amazon’s route to entering, and therefore it would not enter via another
route); and
(b) post-entry unilateral effects, where Amazon re-enters the market
notwithstanding its investment in Deliveroo, but it either (i) competes less
strongly to internalise 16% of Deliveroo’s profits; or (ii) influences
Deliveroo to compete less strongly against it.
Amazon relying on Deliveroo for its presence in online restaurant platforms
69. We have considered first the effect that a 16% shareholding would be
expected to have on Amazon’s incentive to enter the market. The effect of a
16% shareholding is limited in comparison with the effect of a full merger: with
a 16% shareholding, if Amazon enters the market then, for every sale
Amazon wins from Deliveroo, it loses only the 16% share in the profit
19
Deliveroo would have gained from that sale. If Amazon wholly owned
Deliveroo, for every sale it won from Deliveroo, it would lose the full benefit
that Deliveroo would have gained from that sale. In our view, if there was a
strong financial incentive for Amazon to re-enter, it is unlikely the 16%
shareholding in Deliveroo would materially reduce Amazon’s incentive to re-
enter the supply of online restaurant platforms in the UK.
70. As described above, we believe the evidence shows that Amazon views
restaurant delivery as part of an international wider food strategy, [].
Amazon is already developing certain logistics technology in its [] business
similar to that used in restaurant delivery and developing an []. Therefore,
any business decision Amazon takes as to whether to re-enter restaurants is
unlikely to be substantially changed by the financial return of the 16%
investment.
71. There is evidence suggesting Deliveroo could be considered Amazon’s ‘foot
in the door’ and route to re-entry in online restaurant platforms. This evidence
may suggest that Amazon would have less incentive and/or intention to invest
in either acquiring or building an alternative online restaurant platform
business following the investment in Deliveroo.
72. Amazon has submitted that its development of restaurant food delivery in
India shows that Amazon is innovating globally and that ‘it is fanciful that a
16% minority investment in a single service in a single country would prevent
this deeply held commitment to innovation’ and application of this innovation
to the UK. We consider that Amazon’s investment in India supports our
conclusion that Amazon may have an interest in pursuing multiple entry
routes into supplying online restaurant platforms.
73. We consider there is mixed evidence on what impact the Transaction has on
whether Amazon would re-enter the supply of online restaurant platforms in
the UK. Based on this evidence, we do not believe it is sufficiently likely that
the investment in Deliveroo would deter re-entry by Amazon if there was a
strong financial incentive for Amazon to re-enter.
Amazon competing less aggressively in online restaurant platforms
74. We also assessed whether the Transaction could lead to a lessening of
competition as a result of Amazon choosing to enter (despite having invested
in Deliveroo) but having less incentive to compete strongly against Deliveroo
because of the Transaction, giving rise to horizontal unilateral effects.
75. If Amazon were to enter the market and then choose to compete less strongly
against Deliveroo, we would expect Amazon to win fewer customers from
20
Deliveroo than if it competed strongly. Amazon would also be likely to win
fewer customers from the other incumbents (unless it could avoid competing
strongly against Deliveroo while competing strongly against the other
incumbents). Amazon would miss out on 100% of the profits from any
customer it failed to win from any of the incumbents by not competing
strongly. In contrast, Amazon would gain only 16% of the profits from any of
those customers, retained by Deliveroo, who would have switched to Amazon
if Amazon had competed strongly for them. It is possible that these customers
would be more profitable overall if retained by Deliveroo than if acquired by
Amazon, eg if Deliveroo has lower marginal costs or if the cost to Amazon of
acquiring customers is high. However, we have no evidence to suggest this
would be the case and on balance we would expect Amazon to have a strong
preference for acquiring a customer (and receiving 100% of the profits from
that customer) over allowing Deliveroo to retain that customer (with Amazon
receiving 16% of profits), especially if by competing less strongly Amazon
would also forego winning customers from the other incumbents in which it
does not have an interest.
76. Even if Amazon was a close competitor to Deliveroo, evidence shows that
Deliveroo, Uber Eats and Just Eat are becoming more similar, both in terms of
the business models (with all three platforms offering marketplace and
logistics-enabled platforms) and the restaurants they target. Therefore, even if
competition was particularly strong between Amazon and Deliveroo, there
would still likely be material diversion to the other players as a result of
Amazon worsening its offer.
77. We are of the view that it is unlikely the 16% investment in Deliveroo would
cause Amazon to compete materially less aggressively if it did re-enter. We
note that this assessment could be different should Amazon acquire a
materially larger shareholding in Deliveroo in future.
Amazon discouraging Deliveroo from competing against Amazon in online
restaurant platforms
78. Finally, we considered whether Amazon could use its material influence to
worsen Deliveroo’s offer and lessen competition in this way. Under this
scenario, Amazon would recoup 100% of any profit that arises from sales that
are diverted from Deliveroo to Amazon as a result of this strategy. This means
Amazon may have a greater incentive to engage in this behaviour as opposed
to worsening its own offering (where it would only recoup 16% of any extra
profit).
79. While we believe the Transaction will confer on Amazon the ability to exert
material influence over Deliveroo, this influence is less than would arise in an
21
acquisition of a controlling interest. This could make it harder for Amazon to
drive Deliveroo to worsen or reduce its offering if Deliveroo saw this as
commercially damaging or preventing it from engaging in strong growth
opportunities.
80. In addition, we saw consistent evidence of strong competition between
Deliveroo and Uber Eats and Just Eat, which would limit the scope for
Deliveroo to worsen its offer to accommodate Amazon.
81. Having considered the level of Amazon’s influence and the incentives of
Deliveroo’s management and shareholders (aside from Amazon), as well as
the specific market conditions, and taking these factors in the round, we are
not satisfied that the Transaction would lead to an SLC as a result of
Deliveroo competing less strongly against Amazon in the supply of online
restaurant platforms in the UK.
Conclusion on supply of online restaurant platforms in the UK
82. We conclude that in the counterfactual Amazon is likely to re-enter the supply
of online restaurant platforms in the UK. We do not, however, find it
sufficiently likely that the Transaction will have such a material impact on
Amazon’s incentives to re-enter, or its approach following re-entry, to result in
a substantial reduction in potential competition on the balance of probabilities.
Therefore, we have concluded that the Transaction may not be expected to
result in an SLC in the market for the supply of online restaurant platforms in
the UK.
Supply of OCG in the UK
83. We have considered whether the Transaction will lead to horizontal unilateral
effects in the supply of OCG in the UK. Unilateral effects can arise in a
horizontal merger where one firm merges with a direct competitor that
provides and/or is expected to provide a competitive constraint. Unilateral
effects are more likely where the merger eliminates a significant competitive
force or where customers have little choice of alternative suppliers.
20
84. We have assessed whether the Transaction would result in an SLC in the
supply of OCG services due to one or more of the following:
(a) Amazon using its influence over Deliveroo to discourage Deliveroo from
competing against Amazon in OCG;
20
MAGs, paragraph 5.4.12.
22
(b) Amazon avoiding competing directly against Deliveroo in OCG, to protect
the value of its investment in Deliveroo; and/or
(c) Amazon relying on Deliveroo to give Amazon a presence in OCG, rather
than developing its own service to compete effectively in OCG.
85. Amazon’s and Deliveroo’s current OCG offerings are differentiated. Amazon
offers a wider selection of products than Deliveroo, supports larger basket
sizes, has prices comparable to those offered in-store by supermarkets, and
offers relatively slower delivery times. Deliveroo provides delivery of OCG
within 30-minutes, but offers a smaller range of products, smaller basket
sizes, and typically has prices that are higher than those offered in-store by
supermarkets.
86. Both Amazon and Deliveroo have incentives to improve their OCG offers as
the market evolves. Amazon has consistently differentiated itself on speed of
delivery. []. Amazon []. Amazon announced in July 2020 that its Amazon
Fresh same day grocery delivery service would be offered to all Prime
members in the UK at no additional charge.
21
In addition, Amazon is rolling
out [], in the US. Deliveroo has set out ambitious plans for OCG delivery.
87. Competitors in OCG include online restaurant delivery providers such as Just
Eat and Uber Eats, traditional grocers and convenience stores such as
Waitrose, Sainsbury’s and Co-op (which has an OCG offering separate from
its partnership with Deliveroo), and grocery delivery specialists such as
Ocado. Almost all larger players in the categories of online delivered
groceries, online restaurant delivery, convenience stores and traditional
groceries are currently active in trialling or offering OCG services with on-
demand delivery of between 30-minutes and two-hours. Most market
participants share an intention to expand their existing offers and address
their specific limitations. In addition, some services already offer a relatively
large range, fast delivery and prices which are competitive with in-store offers.
While the future development of the market remains uncertain, we consider
that on the basis of the evidence other market participants may be well-placed
to compete in OCG provision.
88. The Coronavirus (COVID-19) pandemic has had a profound effect on grocery
delivery in the UK, as well as on the grocery sector generally and the wider
economy. Most OCG providers, including Deliveroo, have expanded their
grocery offers in recent months. It remains to be seen how Coronavirus
(COVID-19) will affect the OCG market in the medium- or longer-term.
21
For instance, Amazon announced in July 2020 that its Amazon Fresh grocery delivery service would now be
available for free for Prime members. See: Amazon Fresh Homepage.
23
Although Deliveroo has expanded its OCG services during the pandemic,
these changes have not brought it into substantially closer competition with
Amazon: Deliveroo continues to deliver smaller baskets with a smaller
average order value.
Amazon discouraging Deliveroo from competing against Amazon in OCG
89. Turning to our first theory of harm, if Amazon decided to expand its presence
in OCG (including on-demand delivery), or if it was concerned about
Deliveroo’s expansion, it could potentially seek to use its material influence
over Deliveroo to discourage Deliveroo from also expanding its own OCG
offer (which would otherwise be in competition with Amazon).
90. While our view is that the Transaction would provide Amazon with the ability
to exercise material influence over the commercial policy of Deliveroo, this is
not the same as an ability to control that policy. In particular, it does not
amount to an ability to drive policy in a direction that other shareholders,
management or the board object to.
91. As such, we cannot assume that Amazon will be able to drive policy in a
direction that would lead to an SLC if that would lead to Deliveroo foregoing a
compelling commercial opportunity. This could make it more difficult for
Amazon to prevent Deliveroo from expanding its OCG business, especially if
Deliveroo saw a strong growth opportunity for OCG.
92. To the extent that Amazon may be able to influence aspects of Deliveroo’s
strategy in ways that would reduce competitive pressure on Amazon, we
would also need to assess whether this would result in an SLC.
93. While Deliveroo is likely to expand its OCG offering, it is not clear whether this
expansion will bring it into closer competition with Amazon. As noted above,
Deliveroo has recently expanded its services during the Coronavirus (COVID-
19) pandemic, launching partnerships with a number of additional grocery
providers, although we understand these to be similar in nature to its existing
grocery partnerships, ie with a limited range of items.
94. Further, as noted in paragraph 87, we consider that on the basis of the
evidence other market participants appear well-placed to compete in OCG
provision.
95. Considering these issues in the round, our view is that the Transaction is
unlikely to lead to an SLC through Amazon discouraging Deliveroo from
competing in online convenience grocery services.
24
Amazon competing less aggressively in OCG services
96. We have considered whether Amazon may compete less aggressively against
Deliveroo in the supply of OCG services because of its minority (16%)
shareholding in Deliveroo. Following the Transaction, if Amazon increased its
prices and some customers responded by switching to Deliveroo, Amazon
would have a 16% share of any increase in Deliveroo’s profits from these
customers. As a result, Amazon may have an incentive to set higher prices
following the Transaction than in the counterfactual. Conversely, if Deliveroo
faced aggressive competition from Amazon this could reduce its profitability,
which would in turn reduce the profitability of Amazon’s 16% shareholding in
Deliveroo. This possibility could lead Amazon to avoid competing directly
against Deliveroo.
97. The question for us is whether a 16% shareholding reduces Amazon’s
incentive sufficiently that it would prevent Amazon from making investments it
otherwise would have made to improve its OCG offering.
98. The magnitude of these effects depends on the future closeness of
competition between the two Parties (in the counterfactual), their respective
profit margins, and competition from other OCG providers (ie the effect would
be weaker if most customers who switched from Amazon switched to other
providers rather than Deliveroo).
99. As explained above with respect to online restaurant platforms, in broad terms
we would expect the 16% holding arising from the Transaction to produce a
weaker price effect on OCG services than might be expected from a full
acquisition.
100. Furthermore, while the future development of the market remains uncertain,
we consider that on the basis of the evidence other market participants
appear well-placed to compete in OCG provision. We consider that this will
further weaken any price effect arising from the Transaction because (a) if
Amazon increase its prices, customers may switch to providers other than
Deliveroo, and (b) even if Amazon were to compete less aggressively against
Deliveroo, Deliveroo would still face competition from these other providers.
101. Considering these issues in the round, our view is that the Transaction is
unlikely to lead to an SLC arising from Amazon avoiding direct competition
against Deliveroo in the provision of OCG.
Amazon relying on Deliveroo for its presence in OCG
102. Finally, we consider whether the Transaction would harm competition by
reducing the incentive for Amazon to invest in competing more effectively as
25
an OCG provider, because the proposed acquisition either secures Amazon
an option to acquire Deliveroo or is a first step towards a full acquisition.
103. If the market evolved towards faster delivery and Amazon did not have a
widely available one-hour or two-hour service, Amazon could miss out on a
significant opportunity for growth and profits (including flywheel benefits).
Absent the proposed Transaction, Amazon may therefore have a strong
incentive to invest in developing point-to-point logistics capabilities or by
otherwise improving the delivery speeds it can offer or other aspects of its
service to ensure that it can remain competitive if faster delivery increases in
importance.
104. Internal documents from Amazon and other Deliveroo shareholders indicate
that they see the Transaction as a potential first step toward a full acquisition
of Deliveroo, rather than a purely financial investment. If Amazon sees the
Transaction as giving it a plan or a way to achieve a stronger presence in
OCG provision (ie doing so via Deliveroo), it may be less likely to have an
incentive to improve its own OCG service.
105. In considering whether the Transaction may reduce Amazon’s incentives to
improve its own OCG services, we note that:
(a) In our assessment, the counterfactual is Amazon’s likely re-entry into the
online restaurants platform market in future (and the Transaction would
not significantly affect its incentives to do so). Amazon’s likely re-entry into
this market would provide a potential route for Amazon to improve its
OCG services, namely by using the same logistics network for its online
restaurant delivery service and for some or all of its OCG services. We
note that both logistics-enabled online restaurants platforms (Deliveroo
and Uber Eats) offer OCG services.
(b) Amazon may in any case prefer to develop its own OCG service given
that in doing so it could potentially build on its experience in developing
such a service in the US, and [].
(c) While the Transaction potentially gives Amazon a route to acquire
Deliveroo, Amazon would face a significant cost in making such an
acquisition – particularly if Deliveroo’s OCG business increases in value.
It would also face risks eg of being outbid by another buyer.
106. This issue needs to be assessed in the context of wider competition in the
OCG services market, and while the future development of the market
remains uncertain, we consider that, on the basis of the evidence, other
market participants appear well-placed to compete in OCG provision.
26
107. In this context, if the Transaction were to reduce Amazon’s incentives to
improve its own OCG services, improvements by its competitors of their OCG
services could nevertheless lead to effective competition. In addition, we note
that if the market grows substantially this makes it more likely that it will be
seen as a compelling commercial opportunity, both by Amazon and its
competitors.
108. Considering these issues in the round, our view is that the Transaction is not
likely to lead to an SLC by removing the strategic benefit to Amazon of
developing its own OCG service.
Additional theories of harm
109. Third parties have raised additional concerns with respect to the Transaction,
including the following:
22
(a) First, that post-Transaction Amazon would have an incentive to promote
its own groceries business using the combination of Deliveroo’s existing
and established network, and Amazon’s ecosystem, in particular its
Amazon Prime customer base. It was suggested that Amazon could offer
its Prime customers cheap or free delivery, or preferential delivery timings
on Deliveroo, to drive additional traffic to Amazon and/or Deliveroo and
foreclose competitors.
(b) Second, that post-Transaction Amazon could prevent rival grocery
suppliers from accessing Deliveroo’s logistics network at competitive
prices.
110. In its Phase 1 Decision, the CMA considered whether the Parties could
bundle Amazon Prime and Deliveroo Plus (Deliveroo’s subscription service,
which offers customers free delivery) by offering Prime members a discount
on, or free access to, Deliveroo’s services.
23
The CMA found, on a realistic
prospect basis, that the Parties would have the ability to bundle Deliveroo
Plus and Amazon Prime, and could use this strategy to foreclose Deliveroo’s
competitors. The CMA also found, however, that any incentive to engage in a
foreclosure strategy arising from the Transaction would be insufficient to result
in a realistic prospect of an SLC. We have not seen any additional evidence
on this point in the course of our phase 2 investigation that would change this
conclusion and, therefore, did not re-open this issue.
22
See Amazon/Deliveroo merger inquiry webpage.
23
See Phase 1 Decision, 27 December 2019.
27
111. The concern that Amazon could limit the access of other grocery suppliers to
Deliveroo’s platform is in some respects related to our horizontal unilateral
effects theory of harm with respect to OCG. As set out with respect to the
OCG theory of harm, while we consider that the Transaction is more likely
than not to provide Amazon with the ability to exercise material influence over
the policy of Deliveroo, this is not the same as an ability to control that policy.
In particular, it does not amount to an ability to drive policy in a specific
direction against the objections of other shareholders, management or the
board. Furthermore, any attempt by Amazon to restrict third party access to
Deliveroo’s network would likely be to the advantage of Uber Eats and Just
Eat, and as such, if Deliveroo were to engage in such a strategy, it would risk
strengthening its two key competitors in its core market of online restaurant
platforms (as well as other last-mile delivery providers). We consider that this
effect could weaken Deliveroo’s incentives to engage in foreclosure. In
addition, and as previously noted, while the future development of the market
remains uncertain, we consider that on the basis of the evidence, other
market participants appear well-placed to compete in OCG provision.
28
Findings
1. The reference
1.1 On 27 December 2019, the Competition and Markets Authority (CMA), in
exercise of its duty under section 33(1) of the Enterprise Act 2002 (the Act),
referred the anticipated acquisition by Amazon.com NV Investment Holdings
LLC, a wholly-owned subsidiary of Amazon.com, Inc. (Amazon) of certain
rights and a 16% minority shareholding in Roofoods Ltd (Deliveroo) (the
Transaction) for further investigation and report by a group of independent
panel members (the Inquiry Group).
24
1.2 In exercise of its duty under section 36(1) of the Act, the Inquiry Group must
decide:
(a) whether arrangements are in progress or in contemplation which, if
carried into effect, will result in the creation of a relevant merger situation;
and
(b) if so, whether the creation of that situation may be expected to result in a
substantial lessening of competition (SLC) within any market or markets in
the United Kingdom (UK) for goods or services.
1.3 An SLC occurs when rivalry is substantially less intense after a merger than
would otherwise have been the case, resulting in a worse outcome for
customers (through, for example, higher prices, reduced quality or reduced
choice).
25
The CMA only has the power to impose remedies as part of a
merger review where it concludes that a transaction will result in an SLC.
1.4 Having decided to extend the statutory timetable by eight weeks, the Inquiry
Group is required to publish its final report by 6 August 2020.
26
Information on
the conduct of the inquiry, along with the terms of reference, are set out in
Appendix A.
1.5 This document, together with its appendices, constitutes the Inquiry Group’s
findings. Further information relevant to this inquiry, including submissions
from Amazon and Deliveroo (together ‘the Parties’) and from third parties, can
be found on the CMA’s website.
27
24
Under Schedule 4 to the Enterprise and Regulatory Reform Act 2013.
25
Quick guide to UK merger assessment: CMA18, paragraph 3.1.
26
See Notice of extension, 10 June 2020.
27
See Amazon/Deliveroo merger inquiry webpage.
29
2. Coronavirus (COVID-19) outbreak
2.1 During our review of the Transaction, cases of Coronavirus (COVID-19)
began to appear in the UK and across the world. The first cases of
Coronavirus (COVID-19) were identified in the UK at the end of January 2020.
On 3 March 2020, the UK Government published an action plan for dealing
with Coronavirus (COVID-19) covering scenarios ranging from a milder
pandemic to a severe prolonged pandemic. The situation was declared a
pandemic by the World Health Organisation (WHO) on 11 March 2020.
28
Effect of the Coronavirus (COVID-19) outbreak on the UK economy
2.2 The CMA notified its April Provisional Findings on 16 April 2020.
29
Those
provisional findings were based on an assessment of the effect of the
Coronavirus (COVID-19) pandemic on Deliveroo, and on the market, in March
and early-April 2020.
2.3 In early-March 2020 there were signs of increasing concern about the impact
of Coronavirus (COVID-19) on the UK economy. The FTSE 100 fell by 25% in
the three months to the end of March 2020; this was the largest quarterly drop
in the index since 1987. The Bank of England cut the baseline interest rate
from 0.75% to 0.25% on 11 March 2020.
30
2.4 The UK Government issued its first advice on formal distancing on 12 March
2020 and introduced social distancing guidelines on 16 March 2020. On
23 March 2020, the UK Government introduced measures requiring people to
remain at home, other than for certain limited reasons, and requested that
retail stores selling non-essential goods close for a period of three weeks.
31
Similar lockdowns were imposed in many countries across Europe and
around the globe. Following the introduction of these measures, large
numbers of restaurants, including national chains, took the decision to close
during this period.
2.5 Coronavirus (COVID-19) and the lockdown have had a significant effect on
the UK economy. UK GDP fell by 5.8% in March 2020 and 20.4% in April
2020.
32
Official figures on unemployment for the period following the lockdown
28
See World Health Organisation, 11 March 2020, Virtual press conference on COVID-19.
29
See April Provisional Findings, 16 April 2020.
30
See Bank of England, 11 March 2020, Bank of England measures to respond to the economic shock from
Covid-19.
31
See the GOV.UK website, 23 March 2020, PM address to the nation on coronavirus.
32
Office of National Statistics, 13 May 2020, GDP monthly estimate, UK: March 2020 and 12 June 2020, GDP
monthly estimate, UK: April 2020.
30
are not yet available, however, the UK claimant count
33
level increased by
69.1% in April 2020 from 1.2 million people to almost 2.1 million people.
34
The
claimant count in June 2020 was 2.6 million.
35
In addition, one report indicates
that UK consumer spending fell by 36.5% in April 2020 compared to April
2019, following a 6% drop in March 2020.
36
2.6 In March and April 2020, the situation was highly uncertain and changing on a
day-to-day basis. This uncertainty made it particularly difficult for businesses,
and for the CMA, to forecast how the situation was likely to develop. The UK
Government announced the first steps towards easing the lockdown on
10 May 2020
37
and has continued to gradually ease restrictions. The
approach to exiting lockdown differs across the nations within the UK. The
situation, however, remains uncertain.
Effect of the Coronavirus (COVID-19) outbreak on the restaurant
sector
Effect of Coronavirus (COVID-19) in March 2020
2.7 Prior to the shutdown, there was evidence that the Coronavirus (COVID-19)
outbreak was leading new restaurants to sign up with online restaurant
platforms in an effort to offer delivery during a potential shutdown. There were
reports of thousands of new restaurants registering with platforms.
2.8 Although the UK Government announced relaxations to the planning rules
prior to the lockdown to allow pubs and restaurants to operate as takeaways
during the Coronavirus (COVID-19) outbreak,
38
many pubs and restaurants
including many major restaurant chains shut down, at least for a period of
time, when the lockdown began. Restaurants that closed all, or almost all, of
their outlets included Greggs, KFC, Leon, McDonalds, Nando’s and
Wagamama.
2.9 In the immediate aftermath of the lockdown, there was a substantial decline in
the number of restaurants available to offer online delivery on Deliveroo’s
platform. In the seven days to 24 March 2020, Deliveroo had []% fewer
restaurants available in the UK than during the seven days to 7 March 2020.
33
A statistic that seeks to measure the number of people claiming benefit principally for the reason of being
unemployed.
34
Office of National Statistics, 19 May 2020, Employment in the UK: May 2020.
35
Office of National Statistics, 23 July 2020, Coronavirus (COVID-19) roundup: Economy, business and jobs.
36
Barclays, 11 June 2020, UK Consumer Spending Report (May 2020).
37
On 11 May 2020, the UK Government published a document setting out further details of the phases for lifting
the lockdown restrictions (see ‘Our plan to rebuild: The UK Government’s COVID-19 recovery strategy’).
38
GOV.UK website, 17 March 2020, Government to grant permission for pubs and restaurants to operate as
takeaways as part of coronavirus response.
31
Deliveroo reported greater declines in restaurant availability in other European
countries including France, Italy and Spain. Similarly, Deliveroo’s order
volumes for the seven days to 24 March 2020 were []% lower than its order
volumes for the seven days to 7 March 2020. Deliveroo also reported greater
declines in order volumes in other European countries including France, Italy
and Spain.
Effect of Coronavirus (COVID-19) from mid-April 2020
2.10 A number of restaurants began to re-open for delivery and/or takeaway in
mid-April 2020, including chains such as Burger King, KFC and Pret A
Manger.
39
Restaurant re-openings continued through April and into July 2020
with many chain restaurants gradually re-opening additional outlets. One
report on consumer spending data indicated a year-on-year increase of 24.6%
in online ‘eating and drinking spending in April 2020’.
40
2.11 In the first two weeks after the lockdown in the UK, Deliveroo reported a
significant decline in order volumes. In the week commencing 23 March 2020,
order volumes were []% below the prior year equivalent and []% behind
budget. In the week commencing 30 March 2020, orders were []% behind
the prior year equivalent and []% behind budget. Since then, in the UK,
order volumes partially recovered to the point where they were still behind
budget but ahead of the prior year in April (in April 2020, Deliveroo’s order
volumes were []% higher than in April 2019). Profitability per order however
in April increased []. Deliveroo attributes the former to more people being at
home (because of self-isolation) and therefore larger orders from larger party
sizes (ie whole households) and the latter to a higher mix of independent
restaurants ([]) on its platform. Whilst profitability per order returned to pre-
crisis levels in May and June, Deliveroo saw a significant increase in order
volumes, with order volumes []% and []% above its pre-crisis February
budget forecast for those months.
2.12 Other online restaurant platform suppliers experienced a similar pattern to
Deliveroo of an initial decline in orders in March 2020. Just Eat then
experienced a recovery in April and May. In May, Uber Eats was still reporting
depressed orders but expected a [] recovery in June and July:
(a) Just Eat experienced a []% drop in orders in March 2020 as compared
to March 2019, followed by a []% rise in orders in April 2020 and a
39
BBC, 16 April 2020, Coronavirus: Major UK takeaway chains start to reopen.
40
Barclays, 11 June 2020, UK Consumer Spending Report (May 2020).
32
[]% rise in orders in May 2020 as compared to the same months in the
previous year.
(b) Uber Eats experienced a [] decline in orders in March and April 2020
with average weekly orders approximately []% of the volume in the
weeks prior to lockdown. Uber Eats forecast that its orders would be []
for May 2020, [] in June and July 2020.
Effect of the Coronavirus (COVID-19) outbreak on grocery sector
2.13 The Coronavirus (COVID-19) crisis resulted in a significant increase in
demand for groceries. Reports suggest that grocery sales, both in-store and
online, in the UK increased by over 20% year on year in March 2020.
41
Grocery sales in the UK remained high in subsequent months, increasing by
17.2% year on year in the four weeks to 17 May 2020.
42
The increased
demand has been attributed to consumers increasing purchases in
anticipation of potential restrictions on their movement or potential shortages,
and to consumers preparing more meals at home with the closure of many
restaurants, cafes and school or work canteens.
2.14 The increased demand for groceries included increased demand for online
groceries. Reports indicate that online grocery spending was 13% higher year
on year in the first three weeks of March 2020.
43
Market reports also indicate
that the share of grocery sales made online reached 10% in the four weeks to
18 April 2020, an increase from a share of 7.5% at the end of 2019.
44
The
share of online grocery sales continued to increase reportedly reaching 11.5%
by mid-May 2020.
45
Consumer spending data suggested an increase of
81.9% in online grocery spending in April 2020.
46
Actual sales likely
understate the demand for online groceries during this time period as demand
for at least online scheduled delivery was reported to significantly exceed
capacity.
47
Between March 2020 and July 2020, many grocery retailers
substantially increased their online delivery capacity.
41
Kantar, 31 March 2020, Record grocery sales as shoppers prepare for lockdown, for UK Grocery Market Share
Update, 12 w/e 22 March 2020.
42
Kantar, 27 May 2020, Grocery growth accelerates as lockdown eases, for UK Grocery Market Share Update,
12 w/e 17 May 2020.
43
Kantar, 31 March 2020, Record grocery sales as shoppers prepare for lockdown, for UK Grocery Market Share
Update, 12 w/e 22 March 2020.
44
Retail Times, 28 April 2020, Online reaches 10% of all UK grocery spend amid COVID-19 lockdown, Nielsen
reports.
45
Kantar, 27 May 2020, Grocery growth accelerates as lockdown eases, for UK Grocery Market Share Update,
12 w/e 17 May 2020.
46
Barclays, 11 June 2020, UK Consumer Spending Report (May 2020).
47
Metro, 7 April 2020, Coronavirus UK: What supermarkets have online delivery slots available?.
33
2.15 With the increased demand for online groceries and online convenience
groceries (OCG), Deliveroo has entered into additional trial partnerships or
extensions of trial partnerships with retailers including BP, Holland & Barrett,
M&S, McColls, Morrisons and Paypoint. We have also observed that many of
Deliveroo’s grocery partners appear to be offering a wider range of products
on the platform with a greater focus on categories such as ‘fruit & veg’, ‘fresh
meat’ and ‘tins, jars and packets’.
Longer-term effect of Coronavirus (COVID-19)
2.16 It is not possible to predict with any accuracy what effect the Coronavirus
(COVID-19) pandemic will have on the UK economy, or on the online
restaurant platform and OCG sectors in the medium to long-term. It is
possible that at least some of the recent increases in orders will be sustained
as consumers maintain new shopping habits for both food delivery and online
grocery shopping adopted during the crisis. On the other hand, it is possible
that anticipated restaurant closures and a potential for a longer-term decline in
consumer spending due to job losses or increased economic uncertainty
could lead to lower order levels in the medium- to longer-term. There is also
the possibility that demand in the online restaurant platform and OCG sectors
returns to levels that were forecast prior to the Coronavirus (COVID-19) crisis.
2.17 We have considered the impact of the Coronavirus (COVID-19) pandemic on
our assessment, as far as possible, to consider whether it affects our
conclusions with respect to the Transaction.
3. The Parties, the Transaction and the rationale
3.1 This chapter examines each of the Parties on a pre-Transaction basis
(paragraphs 3.2 to 3.25), the Transaction itself (paragraphs 3.26 to 3.30) and
its rationale (paragraphs 3.31 to 3.55).
The Parties
Amazon
Description and ownership
3.2 Founded in 1994 by current CEO Jeff Bezos, Amazon has grown rapidly to
become one of the world’s largest companies with a market capitalisation of
34
c.$1.52 trillion.
48
In 2019, Amazon had over $280 billion
49
in net sales (up
20% year on year), over $14 billion in operating income
50
and approximately
800,000 employees.
51
‘Amazon.com, Inc.’, the parent company of the group,
is publicly listed on the US Nasdaq Global Select Market (NASDAQ) and
incorporated in Delaware in the United States (US).
3.3 The UK is Amazon’s third largest market in terms of net sales (behind the US
and Germany) with approximately 6% of net sales generated from the UK
(approximately $17.5 billion in 2019).
52
Amazon employs approximately
30,000 people in the UK across its Head Office in London, development
centres in Edinburgh, London and Cambridge, its customer services centre in
Edinburgh, fulfilment and delivery sites across the UK and at its Amazon Web
Services (AWS) UK Data Centre.
53
Amazon operates across a large number
of sectors. Its activities include operating as an online retailer and third-party
marketplace provider, a delivery and logistics network operator, a host of
cloud server space, a book publisher, a producer and online broadcaster of
television and films and a manufacturer of electrical devices such as Amazon
Kindle e-readers and Echo devices.
Overview of main activities
3.4 The Amazon business areas most relevant to the Transaction are described
below.
3.5 Online marketplace. Amazon operates an online retail marketplace
connecting consumers with products from Amazon and third-party sellers. The
majority of physical gross merchandise sales on Amazon (58%) are sold by
third parties.
54
Amazon’s UK-based marketplace is available at
Amazon.co.uk.
3.6 Amazon Logistics. Amazon Logistics is a provider of logistics services
internally, and to third party sellers through Fulfilment by Amazon (FBA) and
Amazon Shipping. Amazon Logistics delivers for Amazon.co.uk, [].
48
As at 28 July 2020.
49
Amazon.com, Inc Q4 2019 Financial Results Conference Call Slides, slide 8. In Q1 2020, Amazon had over
$75 billion in net sales (Amazon.com, Inc Q1 2020 Financial Results Conference Call Slides, slide 7).
50
Amazon.com, Inc Q4 2019 Financial Results Conference Call Slides, slide 10. In Q1 2020, Amazon had
around $4 billion in operating income (Amazon.com, Inc Q1 2020 Financial Results Conference Call Slides,
slide 9).
51
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, page 4.
52
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, page 68.
53
See Amazon.co.uk website.
54
Amazon.com, Inc 2018 Annual Report, page 2 (58% represents gross physical merchandising on the
marketplace from third parties).
35
3.7 Amazon Web Services (AWS). AWS is a cloud computing platform offering a
range of ancillary services including infrastructure technologies such as
compute, storage and database services.
55
Globally AWS accounts for
approximately 12.5%
56
of Amazon’s net sales and was the fastest growing
part of the business in 2019 with net sales up 37% year on year.
57
3.8 Physical Stores. In August 2017, Amazon acquired the Whole Foods Market
business for $13.2 billion.
58
Whole Foods Market is a leading natural and
organic supermarket business with approximately 450 stores across the US
and Canada. It also has a small number of stores in the UK (seven). Whole
Foods Market’s UK subsidiary, Fresh and Wild Ltd, had turnover of £107
million in 2018 and recorded an operating loss of £5 million.
59
Amazon also
operates Amazon Go, a small number of convenience stores in the US selling
a range of snacks, ready to eat meals and groceries. In March 2020 it was
reported that Amazon would open its first Amazon Go store in the UK later in
the year.
60
Overall, physical stores account for approximately 6% of total net
sales. The bulk of the physical stores net sales is from Whole Foods
Market.
61,62
3.9 Consumer subscription services. Amazon also offers consumer
subscription services; the most significant being Amazon Prime. For an
annual or monthly subscription, customers receive free delivery, early access
to discounts and offers and access to and unlimited streaming of thousands of
movies, TV shows and Amazon original content, digital music and e-books. In
the UK, Amazon Prime is available for £7.99 per month or £79 annually.
63
Globally, Amazon now has over 150 million Amazon Prime members.
64
Amazon had [] million paid Prime subscribers in the UK in 2019,
65
and
forecasts this to [].
66
55
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, page 20.
56
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, pages 24 and 68.
57
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, pages 24 and 68.
58
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, page 54.
59
Fresh and Wild Limited Report and Accounts 2018, page 8.
60
See, for example, The Guardian (14 March 2020), Amazon to open store without checkouts in UK.
61
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, page 68.
62
Whole Foods Market, Inc, FORM 10-K, page 16, in year to September 2017 (last year prior to Amazon
acquisition stated that it had total sales of $16 billion).
63
See Amazon.co.uk website listing the benefits of an Amazon Prime subscription.
64
Amazon.com, Inc Q4 results 2019, Earnings Release.
65
Mintel Group (7 March 2019), The Amazon Effect – Nine in ten Brits shop on Amazon (‘…four in ten (39%)
consumers have access to Amazon Prime, with just over a quarter (26%) personally being members and a
further 13% sharing access through someone else’s account. Scaled to a national level, this places Amazon
Prime membership in the UK at around the 15 million mark.’).
66
In 2018, [] million customers ([]% UK households) had an Amazon Prime subscription in the UK; Amazon
expects its Prime use base to [] to [] million in 2022.
36
3.10 Amazon Prime customers are able to use Prime Now, which allows ultrafast
delivery on a variety of (food and non-food) items to certain locations. Amazon
Prime customers in London and certain surrounding areas can also use
AmazonFresh (Amazon recently announced that it would no longer charge
Prime members a fee for AmazonFresh)
67
for grocery delivery. []. Amazon
views Prime Now, in part, as a marketing tool encouraging loyal customers to
continue shopping with them, particularly through the free delivery offered.
68
Overall consumer subscription services account for approximately 7% of total
global revenue.
69
3.11 Amazon Restaurants. Amazon previously offered restaurant food delivery in
both the UK and the US through Amazon Restaurants. It launched in the UK
in September 2016. Amazon decided to stop operating Amazon Restaurants
in the UK in [] and exited in November 2018. In the US, it stopped operating
in June 2019. In 2018, the combined US/UK net revenue of the business was
$[] million.
Investments and acquisitions
Minority investments
3.12 For the year ended 31 December 2019, Amazon had $3.3 billion invested in
minority stakes in public and private companies, of which $679 million related
to publicly listed companies.
70
3.13 Amazon identified to the CMA [] companies in which its initial investment
provided it with an equity stake between 5% and 50% at any point in the last
five years. In several cases, Amazon went on to make further investments
subsequent to its initial investment, although only one of these (Ring Inc) has
led to a full acquisition. Amazon’s proposed investment in Deliveroo is [].
71
Majority shareholdings
3.14 Amazon has previously acquired companies operating in both its own and
adjacent markets. These include: Audible.com (digital audiobooks),
67
See AmazonFresh.
68
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, page 27.
69
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, page 68.
70
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, page 33.
71
The investments identified by Amazon were (i) Hungry Machine, Inc d/b/a LivingSocial, a $[] million
investment for a []% ownership interest in 2010; (ii) Rivian Automotive, Inc, a $[] million investment for a
[]% ownership interest in 2019; (iii) Aurora Innovation, Inc., a $[] million investment for a []% ownership
interest in 2019; (iv) Witzig Advisory Services Private Limited, a $[] million investment for a []% ownership
interest in 2019; (v) Red Seam Holdings LLC, a $[] million investment for a []% ownership interest in 2019;
and (vi) Future Coupons Private Limited, a $[] million investment for a 49% ownership interest in 2019.
37
Goodreads and Shelfari (‘social cataloguing’ online books databases), Twitch
(live video gaming streaming) and Annapurna Labs (microelectronics
provider). Some of Amazon’s most recent acquisitions include: Health
Navigator (2019, symptom checking and triage services), eero Inc (2019,
home wifi), PillPack Inc (2018, online pharmacy), Ring Inc (2018, home
security), Whole Foods Market (2017, groceries) and Souq Group Limited
(2018, online marketplace).
72
Deliveroo
Description and ownership
3.15 Deliveroo is a UK based company founded in 2013 whose main activity is
restaurant delivery. In addition to restaurant delivery, the company operates
cloud kitchens through Deliveroo Editions, delivers convenience groceries
through a partnership with Co-op and certain other retailers, and provides
wholesale and data services to restaurants. Deliveroo’s primary market is the
UK, where it operates in over 200 towns and cities, working with over 30,000
partner restaurants and 25,000 riders and employs over 1,000 people directly.
Deliveroo generated approximately []% (£[] million) of its £476 million
turnover in the UK in 2018.
73
3.16 Deliveroo is also active in Australia, Belgium, France, Hong Kong, Ireland,
Italy, Kuwait, Netherlands, Singapore, Spain and the United Arab Emirates.
74
Whilst the company currently generates a significant amount of revenue, the
costs of growing the business and developing new technology mean that it
currently makes a significant loss (£232 million
75
in 2018 after tax), has
negative operating cashflow (£139 million in 2018
76
) and its continued
operation is dependent on periodically raising capital from investors.
3.17 Deliveroo’s ultimate parent company is Roofoods Ltd, a UK registered Private
Limited Company. The main shareholders are outlined in Table 3.1 below.
72
Amazon.com, Inc FORM 10-K, filed on 31 January 2020, page 54.
73
Roofoods Ltd Annual Report and Financial Statements for the year ended 31 December 2018, page 13.
74
See Deliveroo.co.uk website, ‘About Deliveroo’ webpage.
75
Roofoods Ltd Annual Report and Financial Statements for the year ended 31 December 2018, pages 7, 13
and 22.
76
Roofoods Ltd Annual Report and Financial Statements for the year ended 31 December 2018, page 55.
38
Table 3.1: Deliveroo’s major shareholders
Shareholder
Current
shareholding
Shareholding
post transaction
Description
Accel London Management
Limited (Accel)#
[
]%
[
]%
US-based Tech Venture Capitalist fund that
partners with tech founders from inception
through all stages of growth. Other
investments: Facebook, Dropbox, Blablacar,
Vinted, Docolib. Most active European unicorn
investor by portfolio count (eleven).
Bridgepoint Capital Group Ltd
(Bridgepoint)¶
[]%
[]%
Private equity group with €20 billion AUM.
Focus is investing in established market-
leading companies valued at up to €1 billion,
with investments typically ranging from €75–
400 million across six sectors. States that it
serves on boards and seeks to add value by
contributing to strategic development.
DST Global V, L.P. (DST
Global)*
[
]%
[
]%
Large tech Venture Capitalist Fund,
headquartered in Hong Kong, focused on
making late stage investments in the fastest
growing tech companies. Portfolio has
included: Facebook, Twitter, Airbnb, Spotify,
WhatsApp, Alibaba and others including nine
European unicorns.
Fidelity Management and
Research Company LLC
(Fidelity)§
[
]%
[
]%
US-based global Asset Manager with AUM of
$3.2 trillion. Equities division has $1.7 trillion
AUM. Adopts active management strategy with
180+ equity research professionals picking
stocks considering industry, competitors and
macro- economics.
General Catalyst Group
(General Catalyst)^
[
]%
[
]%
US-based Venture Capitalist fund that makes
early stage and growth investments. Portfolio
includes Airbnb, Oscar, Stripe and Gusto.
Greenoaks Capital Partners LLC
(Greenoaks)
[
]%
[
]%
US-based global internet investment firm
focused on ‘concentrated long-term
investments in technology-enabled business
globally’.
Index Ventures (UK) LLP (Index
Ventures)†
[
]%
[
]%
Dual London/San Francisco-based Tech
Venture Capitalist Fund. Significant experience
in funding European tech unicorns with eight in
its portfolio. Also states it provides support and
experience to management. Investments
include Dropbox, Ayden, Robinhood.
T. Rowe Price Group, Inc
(T. Rowe Price)‡
[
]%
[
]%
US-based Global Asset Manager with
$1.2 trillion AUM.
Will Shu (founder)
[
]%
[
]%
Founder and CEO.
Source: CMA analysis of data from Deliveroo.
Notes: A ‘unicorn’ is a tech start-up company that reaches a valuation of $1 billion or more. AUM = assets under management.
#See Accel website and Crunchbase (27 January 2020), European Unicorns Break Out In 2019.
¶See Bridgepoint website.
*See Crunchbase website: DST Global.
§See Fidelity Investments (31 December 2019), Equity Factsheet.
^ See General Catalyst website.
†See Index Ventures website and Wired (15 October 2018), How Index Ventures became Europe's start-up success factory.
‡See T. Rowe Price website.
Overview of main activities
3.18 Deliveroo’s main activities are described below.
3.19 Online restaurant platform. Deliveroo operates both an online food ordering
marketplace and an online logistics-enabled food marketplace. Approximately
39
[]% of orders are delivered by Deliveroo, with []% being ‘marketplace
only’ orders where delivery is carried out by the restaurant. Restaurants are
charged a delivery fee by Deliveroo based on the order value; this fee is
typically [] where restaurants use the online logistics-enabled marketplace.
3.20 Consumers, using either Deliveroo’s website or app, can select from a wide
range of restaurants in their area. In addition to criteria such as price and food
type, customers are shown a restaurant rating based on previous customer
reviews, and an estimated delivery time (in ten-minute ranges). Consumers
are charged for the food ordered and are also charged a delivery fee.
Deliveroo offers a similar service for corporate customers, Deliveroo for
Business, that allows corporate customers to order in catering or to pay for
this service for employees.
3.21 Delivery is carried out by self-employed bicycle, scooter and car riders/drivers
with an average delivery time of 32 minutes from order.
77
[]% of deliveries
are carried out on either scooters or bikes. Riders/drivers are paid on a per-
delivery basis.
78
3.22 Deliveroo’s systems rely on its logistics platform, ‘Frank’.
79
Frank uses AI,
Deliveroo’s historic data, and external data to allocate riders efficiently to each
delivery, with the aim of minimising the average delivery time. When a
customer is browsing for food, Frank calculates an estimated time for delivery.
3.23 Deliveroo Editions. Deliveroo Editions are delivery-only kitchens owned by
Deliveroo and occupied by restaurants to cook orders solely for delivery (by
Deliveroo riders). There are approximately 16 Deliveroo Editions sites across
the UK and approximately 100 restaurants are set up across these sites. Each
restaurant has its own staff, but shares communal facilities such as food
storage. Deliveroo pays for the build and upkeep of the facilities and provides
all kitchen equipment. Restaurants do not pay a rent to occupy the Editions
kitchens and instead are charged commission on orders generated.
80
3.24 Restaurant Services. Deliveroo provides wholesale and data analytics
services to restaurants. Deliveroo will negotiate bulk discounts for food and
packaging which it will then pass onto restaurants, estimating small
restaurants can save c.[]% costs from these services. Deliveroo also offers
access to its data and analytics tools as a performance management tool for
77
See Deliveroo.co.uk website, ‘FAQs’ webpage.
78
See Roocommunity.com website, ‘Applicant FAQs’ webpage.
79
See: Deliveroo.co.uk website, ‘About us’ webpage and Roocommunity.com website, ‘How and why am I
offered specific orders?’ webpage.
80
See: BBC (23 April 2019), Does your dinner come from a ‘dark kitchen’?.
40
restaurants. Restaurants can, for example, track their most popular items on
Deliveroo, access online marketing tools and keep track of their invoices.
81
3.25 Online Convenience Groceries (OCG). Deliveroo delivers convenience
groceries through partnerships with grocers such as the Co-op. This works in
the same way as restaurant delivery, with customers ordering online or on the
app, the grocer packaging the goods and then the rider delivering them. As
with restaurant food delivery, customers are given an estimated delivery time
within a ten-minute window prior to ordering and the food will be delivered in
an average of 32 minutes.
82
Through the partnership with Co-op, Deliveroo
offers a range of approximately [] different products.
The Transaction
3.26 The Transaction comprises [] which Amazon entered into with Deliveroo on
16 May 2019:
83
(a) []; and
(b) [].
3.27 [] conferred on Amazon the following rights over Deliveroo:
(a) []% shareholding in Deliveroo;
(b) the right to appoint one director and one board observer to Deliveroo’s
board;
(c) []; and
(d) [].
84,85
3.28 [] envisages the acquisition by Amazon of a []% shareholding in
Deliveroo (and will confer on Amazon certain minority shareholder rights),
86
[].
87
81
Deliveroo Foodscene UK (3 October 2018), See how you’re performing with us anytime, anywhere.
82
See Deliveroo.co.uk website, ‘FAQs’ webpage.
83
[]
84
[]
85
[]
86
The Parties provided an updated estimate of Amazon’s shareholding post-transaction of 16.03%. This includes
vesting of options and assumes a further allotment of shares in Deliveroo to an individual, []. The Parties did
not provide supporting evidence for this revised figure, details of how it would affect other shareholders’ stakes or
evidence to support the assumption that these shares would be allotted and issued to []. In any event, we do
not consider that this minor change in shareholding would affect our assessment of the Transaction.
87
The German authorities cleared this Transaction on 11 July 2019.
41
3.29 The Parties submitted that [].
88
However, we observed evidence indicating
that []. In particular, we note that []. The Parties’ internal documents also
indicated that [].
89
3.30 We therefore consider that [].
90
[], we have treated the Transaction as an
anticipated merger.
The rationale
Amazon’s rationale
3.31 Amazon submitted that it believes Deliveroo will be a valuable financial
investment for Amazon, given its rapid growth, customer popularity and the
strength of its management team. Amazon believes that food delivery is a
rapidly growing and valuable sector that offers a large and growing
opportunity, and [].
3.32 Amazon’s internal documents provide further context regarding its rationale,
and indicate that its reasons for investing in Deliveroo were not merely
financial. For example, the email seeking approval for the investment from the
Amazon CEO and CFO, which mentions restaurant delivery [], states:
‘[]’.
3.33 Amazon told the CMA that this text had been drafted when Amazon was
considering [] and had been copied into a subsequent email relating to the
investment decision, without amendment. Amazon said it was not, however,
relevant to the investment decision. We do not consider that this explanation
is credible. The document in question was a short email sent to the CEO and
CFO of Amazon seeking approval for a significant investment (and may
therefore reasonably be assumed to have been carefully drafted and
reviewed). Moreover, parts of this text have been updated [].
91
We do not
therefore accept Amazon’s argument that these statements should be
disregarded as ‘cut and paste’ errors.
3.34 Amazon also contended that there is a distinction between something being
[] to its commercial strategy. Amazon submitted that ‘[]’. It further stated
88
[]
89
For example, in a briefing note to Mr Bezos on the proposed investment, Amazon stated ‘[]’.
90
[], the CMA would treat them as successive events within a two-year period having occurred or occurring
simultaneously on the date of the last transaction under section 29 of the Act. In giving effect to this provision, the
CMA may take into account transactions in contemplation (that is where the last of the events has not yet
occurred). (Mergers: Guidance on the CMA’s jurisdiction and procedure (CMA2), paragraph 4.33).
91
[]
42
that ‘[]’. Amazon submitted that although it has ‘[]’, the Transaction is not
[], and [].
3.35 Amazon has also submitted that the statements in the email sent to its CEO
and CFO seeking approval for the investment in Deliveroo are contradicted by
statements in earlier emails from Amazon’s Senior Vice President of Physical
Stores, [], and Amazon’s Senior Vice President of Worldwide Operations,
[] who indicated that they did not support investing in a restaurants
business.
3.36 Although [] and [] indicated that they do not support the investment in
Deliveroo, the documents cited by Amazon suggest that [], had a different
view.
92
In addition, in a separate email chain that did not include [] and []
around the same time, [] (Amazon VP of Worldwide Corporate
Development) wrote ‘[]’. Amazon submitted that [], and that this decision
[] Deliveroo demonstrates that [] decided that Deliveroo []. However,
following Amazon’s decision [] Deliveroo, and the recommendations of []
and [], Amazon articulated the rationale discussed in paragraph 3.32 and
decided to close the Amazon Restaurants business [] it decided to invest in
Deliveroo.
3.37 The potential strategic value of Deliveroo to Amazon is further supported by
Amazon internal documents where Amazon explored [] before it decided to
invest instead. In one document, Amazon discusses potential ‘flywheel’
benefits
93
to Amazon []. Amazon told us that it decided [].
94
3.38 Amazon further submitted that its decision [] is the best evidence that
Amazon believes Deliveroo is not strategically important to Amazon. We
consider, however, that this position is not fully consistent with all of the
evidence we have reviewed, some of which suggests that Amazon may well
([]) decide to pursue the full acquisition of Deliveroo in future. In particular:
(a) An email exchange involving Amazon’s corporate development team
discussing whether Amazon would be interested in a minority investment
in Deliveroo on [].
(b) [] told us that making an investment in Deliveroo was not a ‘[]’.
92
In the email chain referenced by Amazon, while noting that he agreed with [] assessment, [] wrote ‘[]’.
93
The term ‘flywheel’ is used at Amazon to describe something similar to a virtuous cycle, which powers the
business. For example, lower prices lead to more customer visits, more customer visits increase the volume of
sales, and that results in more commission-paying third-party sellers to the site. [].
94
Amazon’s valuation of Deliveroo forecast [].
43
3.39 The potential strategic value of Deliveroo to Amazon is consistent with
Amazon’s broader commercial strategy, as evidenced by Amazon India’s
launch of Prime Food, a restaurant delivery pilot project. [].
95
3.40 We consider the evidence (in particular, Amazon’s internal documents
discussing the Transaction and the restaurant food delivery industry)
demonstrates that Amazon considers Deliveroo to be of potential strategic
value to it in future ([]), and that Amazon’s investment in Deliveroo was
strategic, not merely financial.
Deliveroo’s rationale
3.41 Deliveroo submitted that it needs investment in order to compete against well-
resourced competitors, such as Uber Eats and Just Eat.
3.42 In addition to the Series G funding round led by Amazon, Deliveroo discussed
[] proposed alternative funding structures led by [].
Funding option led by [
]
3.43 The [] funding led by [] envisaged [] of $[] million. Discussions on
this were held in early 2019 leading to a term sheet being circulated in early
2019. However, Deliveroo has stated that negotiations were broken off when
it became clear that the Amazon investment was highly likely to progress in
April 2019. Deliveroo told us that it received oral commitments for the []
$[] million []. It has also submitted that the terms of the investment were
[].
3.44 This funding round valued Deliveroo at $[] billion. Deliveroo has stated that
the benefits of this transaction were that:
(a) [] than that under the Amazon investment; and
(b) [] would be acceptable to the board.
3.45 However, compared to the Transaction, Deliveroo stated it was less attractive
because:
(a) the lower amount available would [];
(b) []; and
95
[]
44
(c) the advantages of Amazon as an investor would be lost.
Funding option led by [
]
3.46 [] aimed to include over $[] million in funding, []. Negotiations between
Deliveroo and [] took place between March and April 2019 with a heads of
terms document in circulation that remained subject to negotiation. As with the
[] funding, negotiations were broken off in late April 2019 with the execution
of the letter of intent with Amazon. Deliveroo has stated that [] had only
committed $[] million at this point [].
3.47 [], through which [] was intending to invest in Deliveroo, is a private fund
[], with capital commitments of approximately $[] million. The fund is not
able to invest more than []% of committed capital in a single portfolio
company. [].
3.48 Whilst [] anticipated that it would have been able to complete the
investment at the time, this was not certain and [] might only have been in a
position to complete a portion of the funding.
3.49 []
3.50 Deliveroo considered that this funding option was unattractive as:
(a) there was considerable execution risk as [] had only committed
$[] million of funding. [];
(b) [];
(c) []; and
(d) [].
Series G funding option led by Amazon
3.51 Deliveroo contacted Amazon in [] ([]) to determine Amazon’s interest in
Deliveroo. From December 2018 to March 2019 Amazon and Deliveroo held
preliminary discussions []. This involved Amazon and Deliveroo entering
into [] and Amazon requesting information from Deliveroo. Amazon
subsequently informed Deliveroo in [] that it was interested in a minority
investment in Deliveroo, leading a Series G funding round. Investment
agreements were subsequently signed on 16 May 2019.
3.52 The Series G funding provided Deliveroo with:
(a) []; and
45
(b) [].
3.53 This funding round valued Deliveroo at $[] million.
3.54 Deliveroo told us that it accepted Amazon’s offer because:
(a) it offered a higher level of funding [];
(b) Amazon is known as a patient investor and shares a similar consumer
focus to Deliveroo;
(c) [];
(d) it involved a smaller number of investors than other available offers; and
(e) [].
3.55 Deliveroo’s internal documents support this rationale, with one stating that
Deliveroo was ‘[]’, and another stating ‘[]’.
4. Jurisdiction
4.1 An anticipated merger must meet the following two criteria to constitute a
‘relevant merger situation’ for the purposes of the Act:
96
(a) First, the arrangements in progress or in contemplation will, if carried into
effect, lead to enterprises ceasing to be distinct.
(b) Second, either:
(i) the UK turnover associated with the enterprise which is being
acquired exceeds £70 million (known as ‘the turnover test’); or
(ii) the enterprises which cease to be distinct supply or acquire goods or
services of any description and, after the merger, will together supply
or acquire at least 25% of all those particular goods or services of that
kind supplied in the UK or in a substantial part of it. The merger must
also result in an increment to the share of supply or acquisition (‘the
share of supply test’).
4.2 The turnover of Deliveroo in the UK in the financial year preceding the date of
the reference to phase 2 exceeded £70 million (approximately £[] million)
for the purposes of section 23(1)(b) of the Act. The remainder of this chapter
therefore, focuses on the first limb of the relevant merger situation test from
96
The Act, section 23. See also CMA2, paragraph 4.3.
46
section 23(1)(a) of the Act, namely the concept of enterprises ceasing to be
distinct.
Enterprises ceasing to be distinct
Arrangements are in progress or in contemplation
4.3 As discussed in paragraph 3.26, on 16 May 2019, Amazon entered into []
97
with Deliveroo:
(a) [], which gives Amazon certain rights over Deliveroo, including board
representation, a []% shareholding ([]); and
(b) [], through which Amazon will acquire a []% shareholding (and
certain other rights) in Deliveroo, [].
98
4.4 Accordingly, we consider that arrangements are in progress or contemplation
within the meaning of section 36(1)(a) of the Act.
4.5 We note the Parties’ view that []. The Parties have not substantiated this
view either in phase 1 or as part of their submissions at phase 2. As
discussed in paragraph 3.29, in our view, there is sufficient evidence to
conclude that [].
Enterprises
4.6 Section 129(1) of the Act defines an ‘enterprise’ as ‘the activities or part of the
activities of a business’. A ‘business’ ‘includes a professional practice and
includes any other undertaking which is carried on for gain or reward or which
is an undertaking in the course of which goods or services are supplied
otherwise than free of charge’.
4.7 A company that owns a business operating as a going concern with the
necessary assets, employees and customer contracts would satisfy the
factual requirements of ‘an enterprise’ for the purposes of the Act.
4.8 We are satisfied that each of the Parties is a ‘business’ within the meaning of
the Act and that, accordingly, the activities of each of the Parties are
‘enterprises’ for the purposes of the Act.
97
[]
98
The German authorities cleared this transaction on 11 July 2019.
47
Ceasing to be distinct
4.9 Section 26 of the Act provides that any two enterprises cease to be distinct if
they are brought under common ownership or common control. ‘Control’ may
include situations falling short of outright voting control. Section 26 of the Act
distinguishes three levels of interest that can constitute control: (i) material
influence, (ii) de facto control, and (iii) a controlling interest (also known as
‘de jure’, or ‘legal control’).
4.10 De jure control generally means a shareholding of greater than 50%, ie a
majority of the voting rights. De facto control refers to situations where an
entity controls a company’s policy, notwithstanding that it does not hold a
majority of voting rights (eg situations where an entity has, in practice, control
over more than half of the votes actually cast at a shareholder meeting). We
do not consider that the Transaction would give rise to either de jure or
de facto control. We have accordingly focused our assessment on whether
the Transaction would give rise to Amazon acquiring the ability to exercise
material influence over Deliveroo.
4.11 Material influence is described in the Act as being able ‘directly or indirectly …
materially to influence the policy of a body corporate … without having a
controlling interest in that body corporate…’.
99
The CMA’s Jurisdictional
Guidance explains that ‘the policy of the target in this context means the
management of its business, and thus includes the strategic direction of a
company and its ability to define and achieve its commercial objectives’.
100
4.12 The ability materially to influence a target’s policy is not an ability to control it.
In particular, it does not amount to an ability to drive policy in a direction that
other shareholders, management or the board object to. Rather, it is the ability
materially to influence relevant strategic or commercial matters, either
positively (ie by persuading the company to pursue particular courses of
action) or negatively (ie by dissuading the company or its management from
pursuing particular courses of action).
101
4.13 The CMA’s assessment of whether a transaction is likely to result in the ability
to exercise material influence requires a broad, case-by-case analysis of the
99
The Act, section 26(3).
100
CMA2, paragraph 4.14: The CMA need not demonstrate that the acquirer has the ability to influence all
elements of the target’s commercial policy; the ability to influence need arise only in relation to certain strategic
matters (such as a potential sale of the target, new product development, key funding decision etc), and/or over
issues going to the business’ ability to define and/or achieve its commercial objectives.
101
CMA2, paragraph 4.22.
48
overall relationship between the acquirer and a target, and will depend on the
facts and circumstances of each case.
102
4.14 The Parties submitted that the CMA does not have jurisdiction to review the
Transaction because Amazon would not be capable of exercising material
influence over Deliveroo.
103
In response to the April Provisional Findings,
Amazon submitted that it would be in the same position as other investors,
that its right to appoint a board director does not enable it to exercise material
influence and that the rights it would acquire are typical of a minority
shareholding.
104
4.15 The remainder of this chapter describes our application of the test for material
influence in the present case. Applying the framework set out in the CMA’s
Jurisdictional Guidance, this chapter presents the evidence as follows:
(a) Influence through Amazon’s shareholding in Deliveroo.
(b) Influence through Amazon’s right to representation on Deliveroo’s board.
(c) Other sources of influence.
4.16 In keeping with the general principle that the purpose of UK merger control is
to enable the CMA to consider the commercial realities of transactions,
105
we
have then considered the overall effect of the influence Amazon would
acquire via each of these channels. We have had regard to the cumulative
influence afforded by factors that may not necessarily, in themselves, be
sufficient to give rise to material influence to determine whether, when
considered together, they mean Amazon will be more likely than not to be
able to exercise material influence over Deliveroo. In this respect, we disagree
with the Parties’ submission that each potential source of influence must be
considered in isolation and is relevant to the assessment only if it is
individually sufficient to confer material influence.
106
4.17 Consistent with the CMA’s Jurisdictional Guidance,
107
we note the importance
of a case-by-case analysis of the overall relationship between the acquirer
and the target. As such, we have conducted a careful review of the facts and
circumstances of this case without assuming that factors considered relevant
102
CMA2, paragraph 4.15.
103
Parties Initial Submission, 24 December 2019, section 3.
104
Amazon response to the April Provisional Findings, 11 May 2020.
105
See for example, Stagecoach/Mainline (1995), CC Final Report, paragraph 2.34.
106
The Parties’ position on this matter underpins many of the arguments advanced in their response to the
Jurisdiction working paper. Amazon’s response to the April Provisional Findings (11 May 2020) reiterates this
point.
107
CMA2, paragraph 4.15.
49
in previous cases are either necessary or sufficient for a finding of material
influence in this case.
108
4.18 After reviewing all of the evidence, the Parties’ representations, and
considering the commercial reality arising from the structure of the
Transaction and other links between the Parties, we conclude that this
Transaction is more likely than not to confer on Amazon the ability to exercise
material influence over Deliveroo due to the cumulative impact of several
avenues of influence.
Influence through Amazon’s shareholding in Deliveroo
4.19 On completion of [] of the Transaction, Amazon will acquire 16.03%
109
of
Deliveroo’s issued share capital and voting rights ([]).
4.20 Deliveroo’s other major shareholders, and their shareholdings post-
Transaction, are set out in Table 3.1 above.
4.21 At phase 1, the CMA considered only the shareholding to be acquired [].
110
Following the Parties’ submissions to the CMA at phase 2, we consider it
appropriate to consider the impact of the shareholding to be acquired [], on
the basis that:
(a) []
(b) It is the Parties’ intention to [].
111
(c) []
4.22 Accordingly, the [] forms part of the arrangements in progress or
contemplation that constitute the anticipated merger referred to phase 2.
112
108
We note Amazon’s argument in its response to the April Provisional Findings that our approach is
‘unprecedented’ and deviates from the approach in certain previous cases. We consider that each of these cases
also turns on its particular facts and that, in a case-by-case assessment of material influence, the approach to
prior cases can be of only limited precedential value.
109
The Parties provided an updated estimate of Amazon’s shareholding post-transaction of 16.03%. This
includes vesting of options and assumes a further allotment of shares in Deliveroo to []. The Parties did not
provide supporting evidence for this revised figure, details of how it would affect other shareholders’ stakes or
evidence to support the assumption that these shares would be allotted and issued to []. In any event, we do
not consider that this minor change in shareholding would affect our assessment of material influence.
110
Phase 1 Decision, paragraph 43 and footnote 33.
111
The Parties’ submissions are all premised on the acquisition by Amazon of a 16.03% shareholding in
Deliveroo. [].
112
Although we are not required to determine whether material influence is conferred in the context of the [],
we do not consider that our assessment would vary materially, as this would include the right to [].
50
Amazon’s voting rights
4.23 The CMA’s Jurisdictional Guidance highlights that, where an acquirer controls
sufficient votes to block a special resolution (generally, more than 25%), this
will typically give rise to the ability materially to influence a target’s policy.
113
4.24 The 16.03% shareholding to be acquired by Amazon is below the level that
would enable it to block a special resolution.
4.25 In some cases, the distribution of shares among other shareholders and
patterns of attendance at shareholder meetings may mean that shareholdings
of less than 25% confer sufficient voting rights to enable the acquirer to block
a special resolution in practice.
114
4.26 The Parties have submitted that Amazon is not expected to be able to
exercise substantially more than 16.03% of voting rights in practice:
(a) Provided that each of the other eight shareholders holding more than
[]% of Deliveroo’s issued share capital (which would represent more
than []% of voting rights in total) attend Deliveroo’s shareholder
meetings, Amazon would not be able to exercise more than 25% of voting
rights.
(b) In the last three years, shareholder resolutions required for significant
events were passed with at least []% of shareholder votes.
4.27 On this basis, we do not consider that Amazon will have the ability to block
special resolutions at a general meeting simply by exercising the voting rights
attached to its shares.
4.28 However, this does not preclude a finding of material influence. In accordance
with the CMA’s Jurisdictional Guidance, we have also had regard to a number
of other factors that may mean that a shareholding of below 25% will
contribute to an acquirer’s ability to exercise material influence including, in
particular:
(a) The existence of veto and other rights enjoyed by Amazon or other
special provisions in Deliveroo’s constitution; and
(b) The ability to influence other shareholders in matters of policy owing to
Amazon’s particular status and expertise.
113
CMA2, paragraph 4.19.
114
CMA2, paragraph 4.21.
51
Amazon’s rights as a shareholder
4.29 Post-Transaction, Amazon, as [] and holder of a majority of Deliveroo’s
Series G shares,
115
will benefit from certain other rights not enjoyed by
Deliveroo’s other major shareholders. These include:
(a) As a Series G shareholder, Amazon will have a higher liquidation
preference and ranking for the two years following adoption of Deliveroo’s
new Articles of Association. Whilst all holders of Preferred Shares have a
superior ranking to holders of all other classes of shares, for this two year
period, the Series G shareholders additionally and uniquely are entitled,
upon liquidation, to a multiple of 1.5x the Series G cost basis.
116
The
Parties have submitted that this is a ‘standard [protection] []’. We
consider, however, that this explanation fails to reflect the fact that the
multiple is higher than enjoyed by holders of any other series of Preferred
Shares, a fact which was []. In addition, [].
(b) [].
117
The Parties have submitted that [], and therefore is not relevant
to the material influence assessment. However, as discussed below, we
consider Deliveroo’s investors will consider it important to keep their exit
options open and that this right contributes to Amazon’s ability to exert
influence over how a sale might be achieved.
(c) [].
118
The Parties have submitted that this is ‘also a normal minority
protection’ but do not dispute that Amazon would be able to exercise this
[] unilaterally, as holder of the majority of the Series G preferred
shares. Whilst a majority of the holders of each other series of preferred
shares must also agree to these matters, we consider it relevant that
Amazon, alone, will control the majority of Series G shares.
4.30 The Parties have submitted that veto rights are relevant to the assessment of
material influence only where they are equivalent to an ability to block a
special resolution or would enable Amazon to persuade other shareholders
that a particular course of action was in their own best interests, and thereby
block a special resolution in combination with these other shareholders. We
consider that this reflects the Parties’ unduly narrow approach to the analysis
of material influence, which assumes that any individual factor is relevant only
to the extent that it confers material influence in itself.
115
Deliveroo Foodscene UK (17 May 2019), Amazon leads a $575MM financing round in Deliveroo.
116
New Articles of Association, adopted 16 May 2019, Article 5.1.1 (see Resolution of adoption of Articles of
Association (published on Companies House website on 11 July 2019)).
117
[]
118
New Articles of Association, adopted 16 May 2019, Article 6.7 (see Resolution of adoption of Articles of
Association (published on Companies House website on 11 July 2019)).
52
4.31 As discussed at paragraph 4.16 above, we consider that the Parties’
approach is not correct, because it fails properly to take into account the
requirement to consider the cumulative impact of all sources of influence. In
addition, we consider the Parties’ focus on the ability to block special
resolutions as ‘the key consideration in the material influence analysis’ is
misplaced. The CMA’s Jurisdictional Guidance states that this is an example
of a situation where it is likely that material influence will be conferred through
a shareholding. However, this was not intended to be (and is not) an
exhaustive definition of material influence: the test is a purposive one,
intended to determine whether an acquirer can influence the target’s
commercial policy. Blocking special resolutions is just one possible means of
obtaining that influence.
4.32 We note that the Parties and Deliveroo’s other shareholders have
acknowledged that [] of Deliveroo [].
119
We consider that this is a matter
that is relevant to the strategic direction of Deliveroo and that the rights
described above are likely to place Amazon in a position to have an influence
on how this strategy might be achieved, whether directly – by leading
management or other shareholders to give greater weight to Amazon’s
position – or indirectly – [] by an alternative acquiror.
Amazon’s status and expertise
4.33 In addition to the rights described above, we have considered whether the
specific status or expertise of Amazon, and its corresponding influence with
other shareholders, may contribute to or give it an ability to influence
Deliveroo’s policy formation.
120
4.34 Status, in this regard, refers to the perception among other Deliveroo
investors of Amazon’s position as an investor. Expertise refers to Amazon’s
knowledge and experience that is relevant to Deliveroo’s business.
4.35 The Parties have submitted that, for status and expertise to be relevant to the
analysis of material influence, these factors must be:
(a) ‘virtually unrivalled’; and
(b) ‘sufficient to lead to an expectation that Amazon could persuade
Deliveroo’s other shareholders that a particular course of action was in
their own best interests in order for Amazon to use their votes to block a
119
[] and see also the evidence from Deliveroo’s investors referred to in paragraph 4.40 and footnote 122.
120
CMA2, paragraph 4.22.
53
special resolution, or could influence materially policy that would
otherwise be expected to require a special resolution’.
4.36 As noted at paragraphs 4.16 and 4.17 above, we consider the Parties’
position in paragraph 4.35(b) relies on the mistaken premise that a factor is
relevant only where it is sufficient to give rise to material influence itself and
overstates the importance of decisions requiring special resolution. In
addition, we consider that the reference to ‘virtually unrivalled’ expertise is not
a reference to a generally accepted principle or requirement for that particular
degree of status or expertise to be present.
121
That said, as set out below, we
consider there is evidence that Deliveroo and its investors did consider
Amazon’s expertise and status to be particularly significant.
Status
4.37 We consider that Amazon’s status as a strategic investor and potential route
to exit for Deliveroo’s venture capital investors is likely to increase the weight
its views will carry with Deliveroo’s management and other investors and, as
such, contribute to its ability materially to influence Deliveroo’s policy.
4.38 The available evidence demonstrates that Amazon was considered to
represent a possible ([]) route to exit for Deliveroo’s other investors. For
example, following a meeting with [], [] flagged that certain intelligence
should be shared with ‘[] and, separately, with like-minded investors who
you know are supportive of an Uber (or Amazon) exit’.
4.39 Similarly, [] commented on Amazon’s intentions regarding its investment in
Deliveroo: []. [] noted, []. [] also commented that ‘[]’. The Parties
have submitted that these emails are not evidence of Amazon’s intent, but this
misunderstands how we consider this evidence to be relevant to our
assessment of Amazon’s likely influence over Deliveroo’s commercial policy.
The emails are evidence of the impression Deliveroo’s investors had of
Amazon and, in turn, of the weight those investors would be likely to place on
Amazon’s point of view in relation to strategic matters.
4.40 More generally, Deliveroo’s existing shareholders are venture capital firms
and the available evidence indicates that most are likely to be looking for an
121
Instead it reflects the terminology used to reflect the facts in that case and (BskyB v CC (Cases 1095/4/8/08
and 1096/4/8/08) [2008] CAT 25, paragraph 138). The CC’s emphasis was on the expectation that BskyB’s views
were likely ‘to be of particular interest to other shareholders’ (BskyB/ITV (2007), CC Final Report,
paragraph 3.59).
54
opportunity to exit [] ([]).
122
This adds further weight to Amazon’s status
as a strategic investor in Deliveroo
123,124
and we consider it is likely to
enhance Amazon’s ability to influence strategic matters, including how a sale
might be effected.
125
4.41 The Parties have submitted that, []. We consider that the shareholders
would wish to keep their exit options open to optimise timing and financial
return on exit, and would therefore continue to attach importance to Amazon
as a potential exit strategy []. Contrary to the argument advanced by the
Parties, we consider that the fact that there may be another option will not
necessarily diminish other investors’ interest in preserving the possibility of
pursuing an Amazon exit in these circumstances.
126
[].
127
4.42 []. We disagree with the Parties that these concerns about the [] were
expressed only in relation to the near term.
128
4.43 In light of the preferred [], we consider that Amazon’s position as a potential
future purchaser (or funder) of Deliveroo is likely to increase the weight its
views would have with both management and the other investors and, in turn,
increase its ability to influence Deliveroo’s commercial strategy.
4.44 Separately, Amazon’s status may also create a perception, among
Deliveroo’s management, of the need to act with caution to avoid conflict with
Amazon’s offering. This is indicated in a []. The Parties submitted that
Amazon is cited in this document because it is a strong competitor, not in the
context of its investment. However, we note that no other competitor is
mentioned despite supermarkets such as Tesco, Ocado and Sainsbury’s all
having well-developed offerings in this sector. The Parties also submitted that
122
[] described itself as having a typical time to exit of [] years (we note that [] original investment in
Deliveroo was in []); [] told us that its ‘typical investment period is [] years’ which would mean an exit in
[] from Deliveroo; [] told us that it hoped to exit its investment in Deliveroo after an initial public offering (IPO)
in [] years; and [] told us that it was assuming a [] IPO as a first potential exit option. [] indicated it was
expecting an IPO in []; [] also indicated it was expecting an IPO in []; a [] document considers possible
IPO []; and a [] email notes, in relation to interest received from [] in buying Deliveroo shares, ‘I trust the
fundraise discussions that are being run in parallel will access the relevant names in that context, particularly
given previous fundraising round and investor access …Of course, should a corporate deal not materialise or any
fundraise/secondary sale process be short on demand, we could contact them then’.
123
[] commented on Amazon’s strategic interest in Deliveroo: ‘Amazon is pushing harder into food delivery
globally (announced investment in India and Indonesia) and their ultimate intentions for Deliveroo clearly go in
that direction’.
124
Breedon Aggregates/Hope Construction Materials (2016), CMA Phase 1 Decision, paragraphs 40 and 44.
125
BskyB/ITV (2007), CC Final Report, paragraph 3.59.
126
Nor do we agree with the [].
127
See also []: ‘Investor sentiment towards heavily loss-making technology companies has substantially
deteriorated since Deliveroo and Amazon signed the deal given high-profile failed technology investments (eg
WeWork)’; []: ‘[]’; and []: ‘[]’.
128
Whilst Deliveroo’s existing shareholders [], evidence from these shareholders indicates that [] is
uncertain and dependent on multiple factors, including investor sentiment towards ‘loss-making technology
companies’ such as Deliveroo.
55
the document ‘[]’. We note that it is not necessary to find that Deliveroo
would be influenced by Amazon not to consider a given strategy; rather, it is
sufficient that Deliveroo’s appetite for pursuing certain strategies might be
reduced.
129
Expertise
4.45 We consider that Amazon’s knowledge and expertise relevant to the areas in
which Deliveroo operates, relative to that of other investors, is also likely to
contribute to its influence with Deliveroo’s management and other
shareholders.
4.46 The Parties recognise that ‘[]’.
130
They submitted that Amazon []. The
Parties have also submitted that the other shareholders, in particular Will Shu,
but also some of the other investors, who have ‘been shareholders in
Deliveroo for a significant length of time’, have relevant expertise.
4.47 We consider that Amazon has significant direct operational experience in
areas that are highly likely to be relevant to Deliveroo’s business, including
those described in the phase 1 decision.
131
4.48 We consider that this experience is likely to give Amazon more in-depth
knowledge and relevant expertise than Deliveroo’s other shareholders,
notwithstanding their experience as investment managers. We note, also, that
a finding that an acquirer’s specific expertise may affect the weight its views
are given by other shareholders or management does not depend on finding
that the other shareholders have no relevant expertise.
132
4.49 The relevance of Amazon’s knowledge is supported by statements made by
both Deliveroo management and its shareholders, as well as other Deliveroo
internal documents. In particular:
(a) Deliveroo’s management expressed [].
133,134
[].
135
We do not agree
with the Parties’ submission that these requests cannot be evidence that
129
BskyB/ITV (2007), CC Final Report, paragraph 3.46.
130
Parties Initial Submission, 24 December 2019.
131
As set out in paragraph 46 of the Phase 1 Decision, Amazon has expertise in operating online platforms
(including logistics-enabled marketplaces), an ultrafast grocery delivery service, a well-established subscription
service, business-to-business offerings, and has experience in new geographic areas for Deliveroo.
132
British Sky Broadcasting Group plc v Competition Commission (CC) (Cases 1095/4/8/08 and 1096/4/8/08)
[2008] CAT 25, paragraph 138, referencing BskyB/ITV (2007), CC Final Report, paragraph 3.59.
133
The Parties have asserted that [].
134
[]
135
Deliveroo’s submitted the following points: []. The CMA does not consider these factors to alter the
conclusion that [] reflect a widely held view within Deliveroo that Amazon’s commercial and operational
expertise is likely to be helpful to Deliveroo’s business.
56
Amazon has relevant expertise because they were compiled by junior
employees. The Parties have not substantiated their position that the staff
in question were junior (or that these staff had limited insight into or
influence over Deliveroo’s commercial policy). Moreover, while the CMA
will carefully consider the author of an internal document and the purpose
for which it was prepared in considering how much weight to give such
evidence, we consider that even junior employees of Deliveroo would be
aware of the insights that would be valuable to Deliveroo’s business.
(b) [] emails [] also highlight a number of areas he anticipates Amazon’s
investment will assist – including [] and [].
136
(c) In addition, [].
(d) Numerous internal documents distinguish the role that Amazon will play
as a ‘strategic’ investor, distinct from the other shareholders and highlight
the importance of having this type of investor. For instance, one email
from [] comments as follows: ‘We are excited about the Amzn
financing. []’.
137
(e) [] analyst recommendation flags Amazon’s expertise in ‘logistics and
tech’ as a key positive for the investment.
(f) [] internal notes state that Deliveroo ‘[]’.
(g) A [] email notes, ‘[]’.
(h) [] also commented, []: ‘[]’.
4.50 The Parties have submitted that []. In this regard, we note that Mr Shu was
deeply involved in negotiating the Transaction, so we do not consider that his
comments can be cast aside as mere []. Moreover, Mr Shu reiterated his
enthusiasm during the phase 1 Issues Meeting, several months after the
Transaction was announced.
138
Similarly, the comments made by Deliveroo’s
other investors are part of an informed analysis of the Transaction, and there
is no reason to believe that these are not considered comments.
136
Deliveroo has described this email as ‘[]’. Even if this were the case, it demonstrates the value that [] –
considered Amazon would have for Deliveroo.
137
Email from []. Also, an email dated [] from []: ‘[]’. Further, an internal email sent on [] by [],
comments: ‘[]’.
138
See footnote 146.
57
4.51 The discussion of possible future partnerships between Amazon and
Deliveroo (discussed further below) also reflects the direct relevance of
Amazon’s operational expertise to the Deliveroo business.
4.52 Moreover, we note that the Parties’ own submissions recognise ‘[]’.
139
4.53 Overall, we consider that there is a strong body of evidence that Deliveroo’s
management, its other shareholders and its commercial/operations teams
perceive that Amazon is a significant ‘strategic’ investor, as well as a potential
future acquirer (or source of funding). Similarly, we consider there is evidence
Deliveroo and its shareholders perceive that Amazon has directly relevant
expertise, which is pertinent to the weight its views are likely to be given
regarding decisions relating to Deliveroo’s commercial policy and strategy,
further contributing to Amazon’s ability materially to influence Deliveroo’s
strategy.
Influence through Amazon’s right to representation on Deliveroo’s board
4.54 Board representation may give an acquirer the ability to exercise material
influence over a target either alone or in combination with other factors.
140
Influence may arise as a result of the particular status of the board
appointee(s) and their ability to shape decision-making through participation in
and shaping of board discussions.
141
4.55 When assessing the impact of board representation, whether as a free-
standing basis for material influence or as a supporting factor in the context of
a broader relationship including a shareholding, the CMA will review a range
of factors and consider how they contribute to the overall material influence
assessment. These factors might include, for example, the corporate/industry
expertise, experience or incentives of the various members of the board.
142
Board representation need not involve board ‘control’ or even confer specific
veto rights at board level in order to be relevant to the CMA’s material
influence assessment.
139
Parties Initial Submission, paragraph 2.2. The desire to retain this investment which, the Parties have said,
‘[]’ may also affect those investors’ perception of Amazon.
140
CMA2, paragraph 4.23, states that material influence may be found where ‘the acquirer is able materially to
influence the policy of the target entity through board representation. Indeed, board representation alone may
confer material influence’. See also Daily Mail General Holdings Limited/ the trustees of the Iliffe
Settlement/Trinity Mirror plc joint venture (2013), OFT Decision document, paragraph 17: ‘The OFT has
previously placed weight on board representation when deciding on the question of material influence’.
141
Daily Mail General Holdings Limited/ the trustees of the Iliffe Settlement/Trinity Mirror plc joint venture (2013),
OFT Decision document, paragraphs 19–20.
142
CMA2, paragraph 4.24.
58
4.56 In this case, we have considered whether Amazon’s representation on
Deliveroo’s board is a relevant factor in the material influence assessment.
Under the [] ([]) Amazon will have the right to appoint one board director
and one board observer. Amazon’s board director will be one of seven voting
directors. [].
143
4.57 Deliveroo’s board directors and their rights post-Transaction are set out in
Table 4.1 below:
Table 4.1: Deliveroo’s board directors and their rights post-Transaction
Name of director
Company affiliation
Rights
Will Shu (founder)
Deliveroo
[]
Martin Mignot
Index Ventures (UK) LLP
[]
Benjamin Peretz
Greenoaks Capital Partners LLC
[]
Luciana Lixandru
Accel London Management Limited
[]
Antoine Froger
Bridgepoint Capital Group Ltd
[]
Doug Gurr
Amazon
[]
Darrell Cavens
N/A (independent)
[]
Adam Valkin
General Catalyst Group
[]
Source: Deliveroo.
4.58 In our April Provisional Findings, and as noted in Table 4.1 above, Amazon’s
nominated board director is Doug Gurr, who is a senior executive at Amazon
and has been head of Amazon UK since 2016.
144
Amazon has since informed
us that Mr Gurr is anticipating leaving Amazon later in 2020.
145
Assuming
Amazon replaces Mr Gurr on the board when he leaves the company later
this year, we would not expect Amazon’s choice of replacement significantly
to affect our assessment of material influence. Whilst our April Provisional
Findings noted Mr Gurr’s individual expertise, the influence of Amazon’s
appointee was driven in large part by the fact that they will represent Amazon
on the board, and will have access to Amazon’s expertise. Moreover, we have
seen no evidence to suggest that Amazon’s preference to appoint ‘[]’ to the
Deliveroo board has changed.
4.59 The Parties have submitted that Amazon’s board representation is insufficient
to confer material influence because of the size of its representation and the
relative experience of Deliveroo’s other board members. The Parties
submitted that Amazon’s appointee will not be able to override the views of
other directors or influence the board to act in a way that is not in Deliveroo’s
best interests.
143
For completeness, the CMA notes that any voting director will have a veto over written resolutions since they
require unanimous approval. Board resolutions which are passed in person require approval by a majority of
voting directors present at the meeting, which is quorate with four voting directors.
144
Mr Gurr’s experience in groceries retail includes nearly five years as development director at Asda, co-founder
and CEO of an online grocery wholesaler, and head of the retail practice at McKinsey.
145
See for example, Natural History Museum, 18 June 2020, Doug Gurr appointed new Director of the Natural
History Museum.
59
4.60 First, we note that the test for material influence does not require Amazon’s
appointee to be able to override the views of other directors or influence them
to act in a way that is contrary to Deliveroo’s best interests. It is not necessary
to demonstrate that Amazon’s appointee is able to drive policy in a direction
that other shareholders, management or the board object to. Rather, it is
sufficient for the CMA to demonstrate that Amazon’s appointee’s views are
likely to be influential in relation to relevant strategic or commercial matters.
4.61 Second, we note the Parties’ submission that Mr Gurr’s knowledge and
experience is at least comparable to that of the other directors, in particular
Mr Shu and Mr Cavens. It is not necessary to find that Amazon’s appointee’s
knowledge is greater than all other members of Deliveroo’s board for it to be
relevant to Amazon’s ability materially to influence Deliveroo and we do not
dispute that, as founder and CEO of Deliveroo, Mr Shu has very significant
knowledge in restaurant food ordering and delivery services. It is clear,
however, that Mr Shu and other investors have recognised the value of
Amazon’s expertise on the board.
146
4.62 We consider that there is ample scope for Mr Gurr’s knowledge and
experience to be applicable to Deliveroo’s commercial policy and that
Mr Gurr, or Amazon’s next appointed representative would be able to obtain
and relay the knowledge Amazon accrued through its experience in restaurant
delivery and from other relevant parts of its business.
147
4.63 Moreover, it is clear that, in nominating Mr Gurr as its initial representative to
Deliveroo’s board, Amazon was conscious of Mr Gurr’s general experience
and seniority, as evidenced by Amazon’s stated ‘[]’. The CMA considers the
decision to appoint the head of Amazon UK as Amazon’s nominated board
representative is indicative of the importance that Amazon attached to its
investment in Deliveroo. We have not seen any evidence that this has
changed since Mr Gurr’s decision to step down from Amazon UK.
4.64 We expect that the combination of the Amazon appointee’s knowledge
(including Amazon’s broader relevant knowledge), their role as a voting
director, and the fact that they are Amazon’s representative on the board, will
mean that their contributions will carry weight among Deliveroo’s voting
directors, in terms of influencing the outcome of board resolutions and earlier
stage discussions relating to the policy of Deliveroo. As a result, we consider
146
[]; in a [] internal email dated 4 May 2019, [] stated that ‘[]’.
147
As recognised in RWE AG/E.ON SE (2019), CMA Phase 1 Decision, paragraph 38, exiting a market does not
mean that the company ‘will lose all of its expertise and operational knowledge acquired over time’. We note that
the Parties’ submissions on the relevant knowledge of other directors’ stems in part from the experience of the
others within their organisation, and that this knowledge may be transmitted through the relevant affiliated
director.
60
that Amazon’s board representation is a relevant contributing factor to our
overall assessment of material influence (see paragraph 4.16 above).
Other sources of influence
4.65 The CMA may also take into account any other factors that may result in, or
contribute to, an acquirer being able to exercise material influence over a
target’s policy. There is no defined list of factors to which the CMA may, or
may not, have regard.
148
Evidence that a transaction is ‘strategic’ rather than
a mere financial investment may be a relevant factor.
149
4.66 We have considered whether other sources of possible influence might result
in, or contribute to, an ability by Amazon to exercise material influence over
Deliveroo. The following sources have been considered:
(a) current and possible future commercial relationships or partnerships
between Deliveroo and Amazon Web Services (AWS);
(b) the possibility of future commercial arrangements or other partnerships
between Deliveroo and Amazon; and
(c) other commercial/operational discussions between Amazon and
Deliveroo.
The AWS agreement
4.67 In previous cases before the CMA and its predecessor authorities, important
commercial relationships – such as supply agreements – have been
considered relevant to the question of material influence, whether by
themselves or because they signal the potential for deeper collaboration in the
future.
150
4.68 Deliveroo currently has an agreement in place with AWS []. The evidence
the CMA has seen also envisages [].
151
4.69 The Parties have expressed the view that the AWS agreement is an []
agreement [], and that there are a number of alternative suppliers available
148
CMA2, paragraphs 4.26–4.27.
149
Daily Mail General Holdings Limited/ the trustees of the Iliffe Settlement/Trinity Mirror plc joint venture (2013),
OFT Decision document, paragraph 24.
150
See First Milk Ltd/Robert Wiseman Dairies plc (2005), OFT Decision document, paragraphs 5–6: Wiseman
accounted for 10–20% of First Milk’s total raw milk sales and the OFT took account of the fact that the transaction
led to ‘a degree of vertical integration’. The OFT found that it was appropriate to have due regard to the possibility
of potential deeper involvement in future.
151
See footnote 154.
61
to Deliveroo. The Parties have also highlighted that Deliveroo []. Finally, the
Parties have stressed that the AWS agreement was negotiated entirely [].
4.70 We consider the following factors to be relevant to whether the AWS
agreement, or factors surrounding it, might contribute to an ability for Amazon
to exercise material influence.
(a) Amazon’s internal documents suggest that []. Evidence of such an
intention in relation to Deliveroo is reflected in [].
152,153
(b) Since the Parties entered into the Transaction, [] AWS and Deliveroo
[] the terms of their agreement [] and its increased reliance on AWS’
services.
154
(c) The terms of the []. The Parties have not substantiated their argument
that ‘engaging multiple providers or switching provider is readily
achievable’, for example by reference to the time and cost implications of
switching supplier. We consider that the ‘[]’ spend commitment
indicates that it would not be realistic for Deliveroo to engage with other
potential suppliers during the term of the current agreement, as suggested
by the Parties.
4.71 Whilst we note the existence of alternative suppliers, we consider that AWS’
relationship with Deliveroo, as well as the potential for the continued
development of this relationship ([]) may nonetheless provide Amazon with
an additional avenue of influence relating to matters relevant to how Deliveroo
defines and achieves its commercial objectives.
Future commercial arrangements/partnerships
4.72 Deliveroo’s internal documents contain a number of references to the
possibility of future partnerships and collaboration between Amazon and
Deliveroo (in addition to the services provided by AWS).
4.73 For example, one document [].
4.74 Other internal documents created after the Transaction signed refer to
furthering cooperation in other areas, as well as a [] and ‘[]’ with the
teams at Amazon.
152
[]
153
[]
154
One internal Deliveroo document notes that ‘[]’.
62
4.75 One of Deliveroo’s other principal investors also recognised the possibility of
future partnerships between Amazon and Deliveroo, commenting that ‘[]’.
4.76 We consider that a perception of Amazon as a strategic/commercial partner
for Deliveroo, both amongst Deliveroo’s management and commercial teams,
as well as amongst Deliveroo’s other shareholders, is relevant to the weight
that Amazon’s views might be afforded in matters relating to Deliveroo’s policy
– in particular its ability to define and achieve its commercial objectives.
Other commercial/operational discussions
4.77 Where an acquirer is likely to be able to exert influence over a target’s
commercial policy outside of formal board and shareholder meetings, the
CMA can take this into account for the purposes of its assessment of material
influence.
4.78 In addition to the possibility of future commercial arrangements being entered
into between Amazon and Deliveroo, there is also evidence (described in
paragraphs 4.49(a) and (c) above in the context of Amazon’s applicable
expertise) that Deliveroo might seek Amazon’s input on a more informal basis
on commercial and operational matters other than at board/shareholder level.
The CMA understands that, in one of the various product areas covered by
the Deliveroo internal log entitled ‘[]’ (described in paragraph 4.49(a)), [].
4.79 We consider that such [] might provide another avenue for Amazon to exert
influence over decisions relevant to the behaviour of Deliveroo in the
marketplace.
Overall assessment of material influence
4.80 As explained above, ‘an assessment of material influence requires a case-by-
case analysis of the overall relationship between the acquirer and the
target’.
155
The CMA will consider the significance of rights conferred by formal
agreements and structures, as well as the commercial realities of the
relationship between an acquirer and the target.
156
4.81 It is important to consider all of the factors outlined above in the context of the
Transaction, and to consider whether cumulatively they give Amazon the
ability to exercise material influence over Deliveroo’s commercial policy. We
note the Parties’ submission that, in circumstances where each individual
155
CMA2, paragraph 4.15. See also RWE AG/E.ON SE (2019), CMA Phase 1 Decision, paragraph 33.
156
See, for example, Stagecoach Holdings PLC/Mainline Partnership Limited (1995), CC Final Report,
paragraph 2.34.
63
factor may not independently confer material influence, the CMA must show
how the factors combine to collectively establish material influence. We have
shown above and conclude below how the various factors combine to
collectively confer material influence.
4.82 Our view is that the factors set out in the foregoing discussion, taken together,
demonstrate that Amazon will acquire material influence over Deliveroo.
157
4.83 This stems from:
(a) the size of Amazon’s investment (in both absolute terms and relative to
other shareholders)
158
and its associated rights;
(b) a strong body of evidence that Deliveroo’s management, its other
shareholders and its commercial/operations teams perceive that Amazon:
(i) has a special status as a significant ‘strategic’ investor, with various
additional rights, and is a credible potential future acquirer of
Deliveroo (or source of funding);
(ii) has commercial and operational expertise from running an online
business in directly relevant sectors to Deliveroo; and
(iii) is a current and potential future strategic/commercial partner of
Deliveroo.
4.84 Together, we consider these factors will mean Amazon’s views are likely to be
given material weight by both management and other shareholders (and their
appointed directors), such that we conclude that the Transaction will give
Amazon the ability materially to influence decisions related to the policy of
Deliveroo, including the management of its business, the strategic direction of
the company and/or its ability to define and achieve its commercial objectives.
4.85 There is also clear evidence that, in practice, there will be ample opportunities
and avenues through which Amazon might actually exert material influence
over Deliveroo’s commercial strategy – including through its participation in
shareholder and board meetings, regular catch-ups with Deliveroo senior
management (between shareholder/board meetings), as well as additional
contacts between senior members of Amazon and Deliveroo’s
commercial/operational teams.
157
It has not been necessary to conclude as to whether any individual factor would confer material influence on
Amazon given that the combination of factors in the round results in material influence.
158
[]
64
Conclusion on relevant merger situation
4.86 In light of the above assessment, our view is that, on the balance of
probabilities, the Transaction will result in the creation of a relevant merger
situation under the Act.
5. Market definition
Approach
5.1 Market definition provides a framework for assessing the competitive effects
of a merger and involves an element of judgement.
159
The boundaries of the
market do not determine the outcome of the analysis of the competitive
effects of the Merger in any mechanistic way, as it is recognised that there
can be constraints on merging parties from outside the relevant market,
segmentation within the relevant market, or other ways in which some
constraints are more important than others. We take these factors into
account in our competitive assessment.
160
5.2 The relevant product market is identified primarily by considering the degree
of demand-side and, to a lesser degree, supply-side substitution. It is usual to
define markets using the hypothetical monopolist test. This test delineates a
market as a set of substitute products over which a hypothetical monopolist
would find it profitable to impose a small but significant non-transitory increase
in prices (SSNIP). The geographic market is also defined using the framework
of the hypothetical monopolist test.
161
5.3 The starting point for our assessment of the appropriate market definition is
the overlapping products of the Parties, or where non-horizontal effects are
relevant, the product(s) of each Party where the non-horizontal relationship
occurs.
162
As set out below, we have therefore considered the following areas
in turn:
(a) Supply of logistics-enabled restaurant marketplaces in the UK
(paragraphs 5.5 to 5.128).
(b) Supply of online convenience groceries (OCG) in the UK
(paragraphs 5.129 to 5.236).
159
MAGs, paragraphs 5.2.1–5.2.5.
160
MAGs, paragraph 5.2.2.
161
MAGs, paragraphs 5.2.10–5.2.20.
162
MAGs, paragraph 5.2.11.
65
5.4 In the following sections we outline our findings and conclusions on market
definition.
Online restaurant platforms
Introduction and overview
5.5 The Parties overlapped in the supply of online food delivery platforms until
November 2018, when Amazon Restaurants exited the UK. Amazon
Restaurants operated as a logistics-enabled marketplace.
163
In July 2018,
Deliveroo began to list restaurants with their own delivery couriers on its
platform, thus also operating as a logistics-enabled marketplace.
5.6 In the present case, for our starting point we have taken the narrowest market
in which the Parties could overlap if Amazon were to re-enter as a logistics-
enabled restaurant marketplace.
164
5.7 As we explain below, we have found that the relevant product market is the
supply of online restaurant platforms:
(a) this market includes food ordering marketplaces, that do not primarily
provide logistics (such as Just Eat);
165
(b) vertically integrated food chains (VIFCs) who provide their own logistics
(such as Domino’s, Pizza Hut and Papa Johns) are not in the market, but
provide an out of market constraint; and
(c) direct ordering from restaurants other than VIFCs is not in the market, but
such restaurants provide a limited out of market constraint.
5.8 Our view is that some aspects suggest a narrower, local geographic market,
and some suggest a national market. In this case we have primarily assessed
competition on a national basis but have considered local factors as part of
our competitive assessment.
5.9 The following analysis is largely based on evidence collected prior to the
Coronavirus (COVID-19) crisis. As discussed above, the Coronavirus
(COVID-19) crisis has had a substantial impact on the market for the supply of
online restaurant platforms (see paragraphs 2.7 to 2.12). Based on our
understanding of these recent developments, we do not consider they have
163
See paragraph 5.10(b) for our definition of a logistics-enabled restaurant marketplace.
164
MAGs, paragraph 5.2.3.
165
See paragraph 5.10(a) for our definition of a food ordering marketplace.
66
changed the nature of demand for, or supply of, online restaurant platforms to
an extent that would lead us to adopt a different view of the relevant market.
Restaurant platform models
5.10 Online restaurant food delivery services have developed rapidly in the UK
over the past 10 to 15-years. Two models for these services have emerged:
(a) the food ordering marketplace model; and (b) the logistics-enabled
marketplace model. As explained in the CMA’s decision in Just
Eat/Hungryhouse:
(a) Under the food ordering marketplace model ‘takeaway restaurants,
contract with the supplier of the platform to join the platform and have
their menus made accessible to consumers. The supplier’s website and
mobile app allow consumers to: search for local takeaway restaurants;
compare menus, prices and reviews; place orders online and pay online
or by cash on delivery. The online orders are transmitted to and accepted
by takeaway restaurants via proprietary terminals, which send
confirmations to consumers, following which the takeaway restaurants
prepare and deliver the food’.
166
(b) The logistics-enabled marketplace model, referred to as ‘ordering and
logistics specialists’ in the Just Eat/Hungryhouse decision, also provides
access to multiple restaurants and consumers on a single platform but ‘in
addition, the delivery of the food to consumers is integrated into the
platform and riders/couriers are able to identify orders that are ready to be
collected in the vicinity. Because they manage the delivery function
themselves, the ordering and logistics specialists have greater control
over the reliability and speed of food delivery than food ordering
marketplaces’.
167
5.11 The existence of these two models is reflected in the Parties’ internal
documents. For example, [].
5.12 As discussed in paragraphs 5.18 to 5.34 below, we note that the distinction
between food ordering marketplaces and logistics-enabled marketplaces is
becoming less relevant as competitors move towards hybrid models.
166
Just Eat/Hungryhouse, Final Report, 16 November 2017, paragraph 2.17.
167
Just Eat/Hungryhouse, Final Report, 16 November 2017, paragraph 2.21.
67
Product market definition
5.13 As set out in the Merger Assessment Guidelines, the relevant product market
is a set of products that customers consider to be close substitutes, for
example in terms of utility, brand or quality.
168
5.14 We consider the extent of substitutability between logistics-enabled restaurant
marketplaces (our starting point) and the products listed below, in order to
assess whether the market definition should be widened to include:
(a) food ordering marketplaces, which do not primarily provide logistics;
(b) direct ordering from branded VIFCs who provide their own logistics, such
as Domino’s, Pizza Hut and Papa John’s;
(c) direct ordering from restaurants (on their own websites or apps, by
telephone, or in-person) who provide delivery (either using their own
drivers or though partnering with a third-party last-mile logistics specialist);
and
(d) direct ordering from restaurants (on their own websites or apps, by
telephone, or in-person) where the customer collects the food from the
restaurant.
5.15 To inform our analysis, we have reviewed evidence on the consumer
decision-making process. As set out in Appendix B there is some evidence to
indicate that consumers typically decide which cuisine they would like to order
before deciding from which app/website to order.
5.16 We consider competition for both consumers and restaurants. We note that
online restaurant platforms have some properties of two-sided markets
169
.
Two-sided platforms intermediate between distinct and unrelated groups of
customers. The number of customers in each group affects the profitability of
the product, because the value that one group of customers realizes from
using the intermediary depends on the volume of customers from the other
group (‘indirect network effects’).
170
• The Parties’ views
5.17 The Parties submitted that the CMA’s market definition should include (a) food
ordering marketplaces, (b) direct ordering from restaurants (with or without
168
MAGs, from paragraph 5.2.5(a).
169
Or three-sided if you include the market for delivery drivers.
170
MAGs, paragraph 5.2.20. As explained in the Merger Assessment Guidelines, the implementation of the
hypothetical monopolist test may be more complicated when products are two-sided.
68
logistics) and (c) VIFCs. If not included in the CMA’s market definition,
Deliveroo submitted that the competitive assessment should give due weight
to these ‘out of market’ constraints on Deliveroo. We consider their views
further in the following assessment.
Food ordering marketplaces
5.18 We first considered whether the products and services of food ordering
marketplaces, which primarily do not provide logistics (such as Just Eat
171
)
should be included in the same product market as logistics-enabled
marketplaces (such as Deliveroo and Uber Eats). These models are
explained in more detail above.
• The Parties’ views
5.19 The Parties’ submitted that logistics-enabled marketplaces and food ordering
marketplaces should be considered part of the same market, and that the
distinction between marketplace-only players and ordering and logistics
specialists has fallen away in recent years. Deliveroo submitted that:
(a) It faces equally intense competition from food ordering marketplaces as it
does from other logistics-enabled competitors. They offer very similar food
delivery services to restaurants/retailers and consumers and have broadly
similar functionalities.
(b) Just Eat is no longer a pure marketplace and has developed a (strong)
hybrid offering in the UK.
(c) Competition for restaurants has grown more intense in recent years due
to the product offerings of the three largest services (Deliveroo, Just Eat
and Uber Eats) converging.
5.20 Deliveroo also submitted that Just Eat’s recent acquisition of the (market-
leading) Canadian food ordering and logistics platform, Skip the Dishes, is
expected to allow Just Eat to leverage the technology and knowhow of this
business to develop further its own delivery offering.
171
Just Eat’s logistics capabilities are not very developed and currently [].
69
• Evidence from the Parties’ internal documents
5.21 The Parties’ internal documents also indicate that Just Eat (primarily a
marketplace) imposes a competitive constraint on Deliveroo and Uber Eats
(primarily logistics-enabled marketplaces). For example:
(a) An internal Deliveroo presentation on ‘[]’ (from December 2019)
includes a section on ‘[]’. This focuses on Uber Eats and Just Eat as
Deliveroo’s key competitors.
(b) In 2019, Deliveroo commissioned price sensitivity research [].
172
This
indicated that [].
5.22 We also observed evidence provided by Deliveroo that consumers and
restaurants multi-home between the two types of marketplace:
(a) Deliveroo submitted that in December 2019, the proportion of Deliveroo
App users who also used the Just Eat app was 37%.
173
(b) Deliveroo submitted the following table (Table 5.1), based on its internal
analysis, indicating that []% of Deliveroo’s restaurant partners in the UK
also work with Just Eat.
Table 5.1: Deliveroo internal analysis on restaurant multi-homing
(%)
UK
Global
Only work with Deliveroo
[]
[]
Also self-deliver
[]
[]
Also work with Just Eat
[]
[]
Also work with Uber Eats
[]
[]
Source: Deliveroo.
Note: [].
• Evidence from third parties
5.23 Evidence from Just Eat and Uber Eats shows that they see themselves as
primarily competing against Deliveroo and one another. Just Eat told us that it
also sees the telephone (ie direct ordering with restaurants) as a main
competitor.
5.24 Just Eat submitted that:
172
Deliveroo commissioned this research with a view to understanding price sensitivity across the market as a
whole.
[].
173
App Annie data showing in a given month the % of phone users opening the Deliveroo app and the Just Eat
or Uber Eats app.
70
(a) it faces competition from a wide range of players, including direct ordering
from local restaurants, VIFCs and ‘logistics enabled platforms’ such as
Deliveroo and Uber Eats; and
(b) businesses such as Uber Eats and Deliveroo are Just Eat’s main
competitors for delivery service orders in the UK, based on their scale and
coverage.
5.25 Uber Eats submitted that:
(a) its main competitors are Deliveroo and Just Eat;
and
(b) there has been a gradual convergence between the business models of
Just Eat, Deliveroo and Uber Eats – with all three offering a food ordering
service with and without access to a delivery network.
5.26 Internal documents from Just Eat and Uber Eats show that these firms
monitor their performance against, and have ambitions to win market share
from, Deliveroo and each other. For example:
(a) Just Eat’s ‘[]’, produced for a board meeting with senior management in
[] 2018, provides evidence that Just Eat views Uber Eats and Deliveroo
as competitors []. [].
(b) Uber Eats’ ‘[]’ describes Deliveroo and Just Eat as its ‘top competitors’.
It provides an overview of Uber Eats’ view of the competition, focusing on
Deliveroo and Just Eat.
5.27 The following slide from a Just Eat internal document on ‘the future of food
delivery’ also supports the views put forward by the Parties [] (see
Figure 5.1).
Figure 5.1: []
[]
Source: Just Eat.
• Evidence from restaurants
5.28 Restaurants responding to our customer questionnaire told us that they use
food ordering platforms because of their delivery capabilities, which are strong
71
due to their scale and investments in dispatching technology.
174
Several
restaurants highlighted the importance of Deliveroo’s logistics services and
told us that they have not seen Just Eat as suitable for them because it does
not provide, or has not previously provided, its own drivers.
5.29 Branded restaurant chains responding to our customer questionnaire
175
told
us that using online restaurant platforms that provide logistics can have
drawbacks including a lack of control over the quality of the product delivered.
For example, one restaurant chain told us that, where it has no control over
the last mile of the delivery process, consumers get a product that is
noticeably worse than that available in the restaurant.
• Our assessment
5.30 Our conclusion is that food ordering marketplaces and logistics-enabled
marketplaces are sufficiently close substitutes to be considered part of the
same product market.
5.31 The distinction between food ordering marketplaces and logistics-enabled
marketplaces is becoming less relevant as competitors move towards hybrid
models, with Just Eat (which previously operated a marketplace-only model)
now offering logistics, and Deliveroo and Uber Eats (which previously only
listed restaurants for which they also made deliveries) now listing restaurants
on their marketplaces which make their own deliveries:
(a) In 2018, Deliveroo introduced its ‘Marketplace+’ offer, which allows
restaurants who make their own deliveries to use Deliveroo.
176
(b) Just Eat has significantly expanded its delivery capabilities, now offering
delivery in over []% of the addressable UK population, and has
purchased the Canadian player Skip the Dishes, which will allow its
technology to be used to further improve Just Eat’s delivery
capabilities.
177
Just Eat has also been able to work with third party
providers to supplement its delivery capabilities where needed.
174
We note that customer questionnaires were sent to former or potential customers of one or both of Deliveroo
or Amazon, both of which offer or previously offered a logistics enabled service. This is unlikely to be
representative of all restaurants, such as those who have a strong preference for self-delivery.
175
Customer questionnaires were sent to former or potential restaurant customers of one or both of Deliveroo or
Amazon.
176
Deliveroo Newsroom, 12 June 2018, Deliveroo to add thousands of restaurants to platform.
177
Just Eat, 15 December 2016, Acquisition of SkipTheDishes.
72
(c) On 31 January 2020, Just Eat was acquired by Takeaway.com which has
logistics expertise in other countries.
178
We note that this could help
accelerate Just Eat’s development of logistics in UK.
(d) In 2019, Uber Eats has also introduced a marketplace which allows
restaurants to make their own deliveries.
179
(e) All three platforms offer additional ‘value add’ services such as, for
example, assistance with third party providers and other business-to-
business tools.
5.32 We note that there is still a degree of differentiation between the services of
platforms that are primarily food ordering marketplaces (Just Eat) and those
that are primarily logistics-enabled (Deliveroo and Uber Eats). For example:
(a) Just Eat’s logistics capabilities are not very developed and currently [].
Uber Eats’ marketplace offerings represent []. Deliveroo’s marketplace
offerings (ie where restaurants deliver food themselves) represent a []
limited proportion of its business.
(b) Food ordering marketplaces that do not offer logistics provide a somewhat
different offer from a restaurant perspective and so may not be
considered as being close substitutes for all restaurants, in particular
those that do not wish to fulfil their own deliveries. Conversely, logistics-
enabled marketplaces give restaurants less control over the quality of
product delivered and customer service.
(c) Uber Eats told us that, from a consumer perspective, logistics-enabled
marketplaces offer typically shorter delivery times because of the
efficiencies that they are able to create.
5.33 We note that, as submitted by the Parties, the distinction between food
ordering marketplaces and logistics-enabled marketplaces is becoming less
relevant as competitors move towards hybrid models.
5.34 In Just Eat/Hungryhouse, the CMA concluded that the relevant product
market included food ordering marketplaces (Just Eat and Hungryhouse) and
the services of ordering and logistics specialists (principally Deliveroo, Uber
Eats and Amazon Restaurants), together referred to as ‘online food
platforms’.
180
In the present case we refer to these as ‘online restaurant
178
Takeaway.com N.V./Just Eat plc, 23 April 2020, CMA clearance decision.
179
FT.com, 20 February 2019, Uber Eats to cut fees in battle with Deliveroo and Just Eat.
180
Just Eat/Hungryhouse, Final Report, 16 November 2017, paragraph 4.28.
73
platforms’ to avoid confusion with the market for the supply of OCG
(discussed in the second half of this chapter).
Vertically integrated food chains
5.35 We also considered whether the relevant product market should be expanded
to include direct ordering from vertically integrated food chains (VIFCs) who
offer their own delivery, such as Domino’s, Papa John’s and Pizza Hut.
5.36 In Just Eat/Hungryhouse, the CMA found that branded food chains who offer
their own delivery, such as Domino’s, Papa John’s and Pizza Hut, should not
be included in the same market, in part because the narrow range of food
types available means that they would be considered close substitutes by
consumers only in relation to a subset of consumer orders.
181
5.37 In Just Eat/ Hungryhouse, the CMA noted that ‘these chains’ services differ
fundamentally from the services that food ordering marketplaces offer, in that
they do not provide consumers a choice of either restaurants or cuisines’.
While VIFC chains accounted for ‘a significant share of takeaway restaurant
revenue, evidence indicate[d] that consumers [we]re unlikely to view these as
sufficiently close substitutes to the restaurants and services offered by [food]
ordering marketplaces and logistics specialists for them to be included in the
same market’. In particular:
182
(a) The results of the CMA’s econometric analysis for Just Eat/Hungryhouse
did not find evidence that Domino’s (the largest of these chains by
revenue and number of branches) exerted a discernible competitive
constraint on Just Eat or Hungryhouse.
(b) The responses to the CMA’s consumer survey for Just Eat/Hungryhouse
did not point towards these chains being close substitutes.
5.38 We note, however, that the evidence from Just Eat/Hungryhouse referred to
here and elsewhere in this paper is from 2017, and that suppliers of online
restaurant food platforms have evolved significantly since then. There may
also have been some evolution in consumer demand in this fast-moving
market (for example, as businesses develop new ways of engaging with
consumers, or consumers change their cuisine preferences). We have
therefore placed limited weight on this evidence. We consider these findings
in the round alongside other evidence collected as part of this investigation.
181
Just Eat/Hungryhouse, Final Report, 16 November 2017, paragraph 4.27.
182
Just Eat/Hungryhouse, Final Report, 16 November 2017, paragraph 4.27.
74
• The Parties’ views
5.39 The Parties submitted that Deliveroo principally competes in the restaurant
ordering and delivery space against Uber Eats and Just Eat, although it also
faces significant competition from vertically integrated pizza chains.
5.40 The Parties submitted that VIFCs that provide their own logistics, such as
Domino’s, Papa John’s and Pizza Hut, impose a constraint on both the
customer and restaurant side of the service. The Parties told us that ‘[]’.
5.41 The Parties submitted that VIFCs provide a strong competitive constraint [].
They supported this with evidence from [].
5.42 Deliveroo submitted that, for customers wishing to order a pizza, there is very
strong demand-side substitution between its offering and the offerings of
Domino’s, Papa John’s and Pizza Hut. Deliveroo submitted that these players’
offerings are comparable in terms of price, service and quality.
5.43 Deliveroo submitted that a significant proportion of their customers wish to
order pizza:
(a) []
(b) []
5.44 The Parties submitted that market intelligence reports and articles suggest
that there may be increasing competition between VIFCs and online
restaurant platforms.
183
5.45 The Parties submitted that there is a trend of large chains moving towards
vertical integration, citing ‘The Sift Report 2020’.
184
In particular, the Parties
reference a quote from the CEO of Keatz (a European ‘cloud kitchen’) saying
that ‘a lot of big chains are trying to become independent – deliver themselves
or work with [a logistics firm like] Stuart, […] [t]here’s an entire industry
developed around this use case; it’s quite a big threat to Deliveroo and Uber
Eats’. The Parties also referenced text in the report stating that ‘[r]estaurants
of all sizes are choosing to go it (almost) alone. With greater control of their
own orders, brands save on commission fees, build loyalty with customers,
and release. And there’s a startup (or several) to help them with that’.
183
Bloomberg, 15 March 2017, Domino’s, Atoned For its crimes against Pizza and built a $9 billion empire;
Diginomica, 1 October 2018, Why Domino’s Pizza keeps on top of what the competition is up to in digital.
184
The Sifted Intelligence Unit (backed by the Financial Times and sponsored by Uber Eats), January 2020,
The Sift Report 2020: future of on-demand food delivery.
75
5.46 The Parties submitted that if several large chains were to vertically integrate in
the future, in aggregate, Deliveroo would be competing with them for more
than just the subset of consumers who want to order pizza.
• Evidence from the Parties’ internal documents
5.47 As submitted by the Parties, a number of Deliveroo internal documents [].
However, the in the round our view is that Deliveroo’s internal documents
have a clear focus on Uber Eats and Just Eat as its key competitors.
• Evidence from third parties
o Just Eat
5.48 Just Eat submitted that it competes closely with VIFC and explained this is
because of the brand strength of these chains. Just Eat submitted ‘they have
the strength of brand that competes against us’.
5.49 A number of Just Eat internal documents []. However, the Just Eat internal
documents that we have reviewed show a clear focus on Uber Eats and
Deliveroo.
o Uber Eats
5.50 Uber Eats told us that it sees itself as facing a competitive constraint from
VIFCs, as it is responding to the same consumer need.
5.51 It told us that it ranks its competitors in food delivery as [] first, then [],
then ‘other existing players’ (including []).
185
5.52 It told us that it thinks some consumers will consider going to an aggregator
and then make a choice [about which restaurant to order from] from there,
whereas others will consider directly going to some of the brands that they
can go directly to, such as Domino's or Pizza Hut Delivery.
5.53 Uber Eats’ ‘[]’ slide deck presents the findings from its brand tracker survey
of food delivery users. It includes the following slide, [] (see Figure 5.2).
185
Uber Eats also listed, after ‘other existing players’, ‘other potential players’ currently focused on on-demand
groceries (including Amazon Prime Now etc).
76
Figure 5.2: []
[]
Source: Uber Eats.
o Domino’s Pizza
5.54 Domino’s Pizza is not partnered with any online restaurant platforms in the
UK.
5.55 Domino’s submitted that [].
5.56 It submitted that it views Just Eat, Deliveroo, Pizza Hut, Papa John’s, Uber
Eats, McDonald’s, KFC and Burger King as its key UK competitors, or
potential competitors, [].
5.57 It submitted the following slide as part of its explanation [] (see Figure 5.3).
We note that this shows [].
Figure 5.3: []
[]
Source: Domino’s.
5.58 Domino’s told us that it chooses to operate its own order and delivery channel
rather than use online restaurant platforms for a number of reasons. These
include: [].
5.59 However, it also submitted that there are a number of disadvantages to
operating its own order channel. For example, [].
5.60 It submitted that it expects [].
o Papa John’s
5.61 Papa John’s is partnered with Just Eat, Deliveroo and Uber Eats. It submitted
that it lists on these platforms to access a wider customer base, and it expects
the proportion of its customers using aggregators to go up in the next 24-
months.
5.62 It submitted that it had built its brand using its own platform but uses online
restaurant food platforms to increase brand accessibility. It submitted that the
benefit of using its own delivery fleet is that it has control over customer
orders from beginning to end. However, there are increasing labour costs
associated with operating its own delivery channel.
77
5.63 It submitted that it ranks Domino’s and Pizza Hut as its top two key
competitors, or potential competitors, followed by Deliveroo and Uber Eats.
5.64 Papa John’s submitted that ‘3 or 4 years ago you could only order pizza,
Chinese or Indian for home delivery. Now customers can choose whatever
food type they want and this would be delivered. This creates challenge for
pizza industry as there is more choice for customers’.
o Pizza Hut
5.65 Pizza Hut partners, in the UK, with Just Eat, Deliveroo and Uber Eats. It
submitted that the benefit of using aggregators is that they have large media
budgets to attract customer attention and create new customer awareness for
the brand. However, there are disadvantages – there is also little flexibility in
online platform design to promote own brand personality. Sharing and use of
customer data also remains a significant challenge for aggregators and
brands to solve for fairly.
5.66 Pizza Hut told us that it ranks Domino’s followed by Papa John’s as its key
competitors or potential competitors.
• Our assessment
5.67 Online restaurant platforms are a relatively new service. Deliveroo, Just Eat
and Uber Eats have been investing heavily in rapid growth. Internal
documents show they see themselves as being in a race to achieve market
leadership as the market matures. We consider that this dynamic is central to
the business models of the platforms and, because of this, the primary
competitive constraint they face is from one another.
5.68 In contrast, VIFC firms are well-established. While the platforms may win
some business from VIFCs, they are focused on expanding the market
beyond this, for example through expanding to more independent restaurants
and other QSR (quick service restaurants) that do not usually operate their
own delivery channel.
5.69 Besides competing for customers, the online restaurant platforms also
compete against one another to attract restaurants to their apps, and for
couriers. However, VIFCs do not compete against the platforms for
restaurants. While VIFCs may to some extent compete against platforms to
attract couriers, we have focused on consumers, and the extent to which
VIFCs are a competitive constraint on online restaurant platforms in winning
delivery orders. We note that:
78
(a) All of the VIFCs identified as competitors by the Parties are predominantly
pizza delivery companies.
(b) Whether VIFCs compete against online restaurant platforms depends on
whether consumers consider VIFCs as a possible alternative, at the point
when they are ordering from an online restaurant platform:
(i) Consumers
186
may compare food delivery options across a range of
factors, including cuisine type as well as menu prices, speed/reliability
of delivery, delivery charge etc. In this case, VIFC options could be
considered as an alternative to all the cuisine options available from
Deliveroo.
(ii) Alternatively, consumers may choose the cuisine first and then
consider provider/delivery options within this cuisine type. Evidence
indicates that this is the typical approach.
187
If the consumer’s cuisine
choice is pizza, they will be able to compare VIFC options against
only a relatively small subset of options available from Deliveroo. [].
However, our own analysis of Deliveroo’s data on gross
merchandising value for its ‘key account restaurant customers’
indicates []. For example, [].
(c) In other words, there may be limited scope for an online restaurant
platform to ‘win’ a customer who has already decided to order pizza.
Hence the competitive interaction between online restaurant platforms
and VIFCs may be relatively limited. Likewise, consumers who do not
want to order pizza are unlikely to view VIFCs, such as Domino’s, as a
substitute to online restaurant platforms.
5.70 We have considered a number of points raised by the Parties in support of
their view that online restaurant chains face competition from VIFCs.
(a) Deliveroo provided some evidence from internal documents indicating that
it sees VIFCs as a competitive constraint []. However, [], in the round
our view is that Deliveroo has a clear focus on Uber Eats and Just Eat as
its key competitors, and [].
186
In practice the consumer choice may be made by a household rather than an individuals, and the way the
choice is made may differ from one purchasing occasion to another.
187
While decision-making processes can vary, there is some evidence to indicate that consumers typically decide
which cuisine they would like to order before deciding on which app/website to order from (see Appendix B).
79
(b) The Parties cited press articles
188
which refer to increasing competitive
pressure on VIFCs from online restaurant platforms. We note that while
VIFCs may see the rapid growth of online restaurant platforms as a threat
to their established businesses, it does not follow that the online
restaurant platforms face an effective competitive constraint from VIFCs.
(c) The Parties cited a report
189
which comments that ‘restaurants of all sizes
are choosing to go it (almost) alone’ (ie to move away from using online
restaurant platforms) in order to save on commission fees and have
greater control over their orders. However, we note that the report also
discusses the benefits that Deliveroo and Uber Eats can offer restaurants,
such as increased order volumes, data insights, and allowing restaurants
to focus on their food and in store dinners. We recognise that the
constraint on online restaurant platforms from VIFCs could potentially
increase in the future if more large branded chains were to move towards
vertical integration.
5.71 In conclusion, our view is that VIFCs are not an effective competitive
constraint on online restaurant platforms. While the platforms may win some
business from VIFCs, they are focused on expanding the market in
competition with one another.
Other direct ordering from restaurants other than VIFCs (on their own websites or
apps, by telephone, or in-person) who provide their own logistics
5.72 We also considered whether the relevant product market should be expanded
to include direct ordering from restaurants (on their own websites or apps, by
telephone, or in-person) who provide their own logistics (using their own
delivery fleet or via a white-label third-party logistics specialist).
5.73 In Just Eat/Hungryhouse, the CMA found that the constraint from direct
ordering was limited, despite the fact that most of the restaurants using Just
Eat and Hungryhouse had their own delivery capabilities.
190
The CMA
excluded direct ordering from the market, while taking it into account as part
of the competitive assessment.
191
188
Bloomberg, 15 March 2017, Domino’s, Atoned For its crimes against Pizza and built a $9 billion empire;
Diginomica, 1 October 2018, Why Domino’s Pizza keeps on top of what the competition is up to in digital.
189
The Sifted Intelligence Unit (backed by the Financial Times and sponsored by Uber Eats), January 2020,
The Sift Report 2020: future of on-demand food delivery.
190
Just Eat/Hungryhouse, Final Report, 16 November 2017, paragraph 6.143.
191
Just Eat/Hungryhouse, Final Report, 16 November 2017.
80
• The Parties’ views
5.74 The Parties submitted that direct ordering from restaurants (on their own
websites or apps, by telephone, or in-person), who provide their own logistics,
imposes a constraint on both the customer and restaurant side of the service:
‘[]’.
5.75 The Parties submitted that [] shows that []% of Deliveroo’s restaurant
partners also self-deliver.
• Evidence from the Parties’ internal documents
5.76 As outlined in more detail in the previous sections, Deliveroo’s internal
documents focus on Just Eat, Uber Eats, and sometimes VIFCs as its main
‘competitor set’. They do not refer to direct ordering from restaurants as a key
source of competition.
5.77 []
Figure 5.4: []
[]
Source: Deliveroo.
• Evidence from third parties
o Just Eat
5.78 Just Eat submitted that ‘direct ordering (eg over the phone) from local
restaurants and large chains is, and will continue to be, an important
competitive constraint for Just Eat, as virtually all of Just Eat’s independent
restaurants operated in the takeaway market before they joined any platform’.
It submitted that ‘[r]estaurants have a personal relationship with their
customers and this relationship is maintained through direct ordering and
restaurants own delivery service – this can be via walk-in, phone ordering, or
the restaurant building their own website/app for online ordering’.
5.79 It told us that direct ordering is still a huge part of the market, and that white-
label apps are making it easier for restaurants to create their own websites
and apps. Restaurants want to retain control, so they set up their own
websites and apps to try and go direct.
81
5.80 Just Eat told us that it thinks that as a customer you want one app with
hundreds of restaurants, rather than hundreds of apps for hundreds of
restaurants.
5.81 []
o Uber Eats
5.82 An Uber Eats ‘[]’ document indicates that direct ordering from restaurants is
still the most common food delivery service in London, and that Uber Eats
aims to [].
5.83 An Uber Eats ‘[]’ document also describes direct order only users and
online restaurant platform users as [].
• Restaurants
5.84 Branded restaurant chains
192
told us that they used online restaurant
platforms for a number of reasons, including their delivery capabilities and
access to customers they would not otherwise be able to reach. They also
submitted that using online restaurant platforms can have drawbacks,
including high commission rates, a lack of control over the quality of the
delivered product and a lack of access to consumer data.
5.85 We asked branded restaurant chains whether they would consider creating
their own ordering channel. 18 of the 28 restaurant chains that responded to
our phase 2 questionnaire submitted that they had considered or were trialling
their own ordering channel. However, all of these restaurant chains
highlighted the difficulties in doing so. We note that if more restaurants
vertically integrated there may be potential for a future increase in competitive
constraint, but we note the evidence of current difficulties in vertically
integrating.
5.86 A number of Deliveroo’s restaurant customers told us that offering delivery
would be very difficult without the scale that online restaurant platforms have.
• Our assessment
5.87 Our conclusion is that direct ordering from restaurants which provide their own
logistics is not in the market, but that such restaurants are an out of market
192
Customer questionnaires were sent to former or potential restaurant customers of one or both of Deliveroo or
Amazon.
82
constraint. This is consistent with Just Eat/Hungryhouse, in which the CMA
excluded direct ordering from the market, while considering its strength as
part of the competitive assessment.
193
5.88 Overall, direct ordering is likely to be a less significant constraint in this case
compared with Just Eat/Hungryhouse because the majority of Deliveroo’s
restaurants in the UK are not available for direct ordering. Deliveroo, an
ordering and logistics specialist, provides the delivery capability to a large
proportion of its restaurant customers. For example, Deliveroo submitted that
an estimated []% of its restaurant partners were also available for direct
ordering in 2019. In comparison, []% of Just Eat’s orders were fulfilled by
restaurants’ own couriers in 2019.
White label last-mile logistics specialists
5.89 Restaurants are able to utilise third parties to develop both their ecommerce
capabilities and their delivery and logistics under their own branding.
5.90 For restaurants without their own logistics, using a white label solution may be
an alternative option to using a logistically enabled marketplace. We have
therefore considered whether this should be included in the same product
market.
• The Parties’ views
5.91 The Parties submitted that Deliveroo faces significant competition from large
chain restaurants adopting a white-label approach to ordering/delivery.
5.92 The Parties submitted that:
‘From a restaurant perspective, it is increasingly easy for branded
chains to vertically integrate. Very recent sector intelligence
reports [The Sift Report 2020] suggest that restaurants are
increasingly able to “go it alone” (using white label food ordering
services such as Flipdish to accept orders and, for logistics, either
fulfilling deliveries themselves or working with a delivery partner
such as Stuart). Nando’s, for example, in partnership with a third-
party logistics provider, has recently started offering its own food
delivery in certain areas of the UK. KFC, [], is also trialling its
own delivery services’.
193
Just Eat/Hungryhouse, Final Report, 16 November 2017.
83
5.93 The Parties submitted that:
‘[]’.
5.94 The Parties also submitted that Slerp, a UK based white-label platform, has
recently partnered with a number of restaurants and pubs across London. The
Parties submitted that ‘platforms like Slerp are making it even easier for
branded chains in the UK to develop their own on-demand delivery offerings,
and this can only be expected to increase in the future’.
• Evidence from the Parties’ internal documents
5.95 As outlined in more detail in the previous sections, Deliveroo’s internal
documents focus on Just Eat, Uber Eats, and sometimes VIFCs as its main
‘competitor set’. They do not refer to direct ordering from restaurants as a key
source of competitive constraint.
• Evidence from third parties
5.96 Uber Eats submitted that it ‘does not compete for consumers solely with
online restaurant platforms. Indeed, Uber Eats considers that it competes with
a large number of players including restaurants ([]), on-demand grocery
providers ([]) and companies with parcel delivery and logistics capabilities
([])’.
5.97 One restaurant chain told us that it used to outsource its logistics to a white-
label logistics provider (Stuart). However, they could not handle the demand
at peak periods as the logistics provider’s capacity was not sufficient. This
caused reputational damage due to late or undelivered orders as it was under
its own brand white-label delivery. It told us that it is unlikely it would use a
third-party logistics company in the future, due to the lack of scalability it
previously experienced. However, it noted that, if a new operator came along
that allowed it to deliver food at a much more competitive price than it
currently pays its online platform, this could still be an option.
• Our assessment
5.98 Our conclusion is that direct ordering from restaurants using a white-label
logistics specialist is not in the market, but that such services are an out of
market constraint.
5.99 Orders fulfilled using white-label logistics specialists are not in themselves a
distinct product from the view of consumers. From a consumer perspective,
this is the same as restaurants that provide their own logistics.
84
5.100 White-label providers of logistics services are likely to be a substitute for
online restaurant platforms for some restaurants. We note that there are
cases in which business may be diverted away from online restaurant
platforms by restaurant platforms using white-label logistics specialists.
However, using these services will not give them access to the benefits they
get from partnering with an online restaurant platform, such as access to
consumers ordering through online restaurant platforms and a reduced risk of
reputational harm from late orders.
Direct ordering where the restaurant does not provide logistics
5.101 We have also considered whether the relevant product market should be
expanded to include direct ordering from restaurants (on their own websites or
apps, by telephone, or in-person) who do not provide logistics (ie where the
customer collects the food from the restaurant).
• The Parties’ views
5.102 The Parties submitted that direct ordering where the restaurant does not
provide logistics is a competitive constraint:
‘[]’.
• Our assessment
5.103 It is our view that direct ordering from restaurants that do not provide logistics
is not in the market, but such restaurants will provide a limited out of market
constraint.
5.104 Consumers’ willingness to pay for food delivery will be affected by their ability
and/or willingness to collect their orders from the restaurant. However,
ordering restaurant food for delivery is attractive to customers who place a
premium on getting their food delivered, and are therefore willing to pay a
delivery fee and service charge. A very high proportion of Deliveroo, Uber
Eats and Just Eat customers order food for delivery.
194
For this reason, direct
ordering from restaurants that do not have their own logistics is likely to
impose a weaker competitive constraint on online restaurant platforms than
those that do.
194
We note that Deliveroo and Just Eat do now offer a click and collect service.
85
Conclusion on product market
5.105 Our view is therefore that the relevant product market in which to assess the
effect of the transaction is the market for the supply of online restaurant
platforms. This includes logistics-enabled marketplaces (such as Deliveroo
and Uber Eats) and food ordering marketplaces, that do not primarily provide
logistics (such as Just Eat). VIFCs who provide their own logistics (such as
Domino’s, Papa Johns and Pizza Hut) are not in the market but provide an out
of market constraint. Direct ordering from restaurants is not in the market, but
such restaurants provide a limited out of market constraint.
Geographic market definition
5.106 The geographic market is also typically defined using the framework of the
hypothetical monopolist test.
195
In the following section, we have explored
whether evidence indicates the market for online restaurant platforms is local
or national in scope.
• The Parties’ views
5.107 The Parties noted the CMA’s finding in Just Eat/Hungryhouse that the
relevant market for the supply of online food platforms was national in scope,
with important local elements needing to be taken into account in the
competitive assessment.
196
5.108 The Parties submitted that Deliveroo’s business is most developed in London
([]). However, Deliveroo has [].
(a) The Parties submitted that []. The Parties also submitted that Just Eat
has recently partnered with Greggs, a large (and popular) nationwide
chain.
(b) The Parties also submitted that Uber Eats’ geographic profile is similar to
Deliveroo’s: [].
5.109 The Parties submitted that Deliveroo ‘uses a number of metrics to assess the
strength of competitors by geography, including []’.
195
MAGs, paragraph 5.2.21.
196
Just Eat/Hungryhouse, Final Report, 16 November 2017, paragraph 4.33.
86
• Evidence from the Parties’ internal documents
5.110 Overall, Deliveroo’s internal documents indicate that it thinks about its position
in the market and competition at both a UK and a local level.
5.111 Internal documents indicate that Deliveroo has ambitions to expand
nationwide. For example:
(a) A slide that was produced for [].
(b) The following slide (see Figure 5.5) that was produced for []. The
Parties submitted that, to the extent to which its internal documents state
that its ambition is to ‘[]’, this was an []. The Parties submitted that
this proposition relied on [].
Figure 5.5: []
[]
Source: Deliveroo.
5.112 Deliveroo’s internal documents show that []. For example, Deliveroo with
Saturday Night Takeaway and the England Football Team, Uber Eats with
Love Island UK, Just Eat with The X Factor.
5.113 Deliveroo submitted that it ‘uses a number of metrics to assess the strength of
competitors by geography, including []’. This is supported by its internal
documents which indicate that Deliveroo considers competition on a local as
well as national level. Generally, [].
5.114 Deliveroo’s internal documents indicate that []. This indicates that Deliveroo
considers how to compete effectively on the local as well as national level.
• Evidence from third parties
o Just Eat
5.115 Just Eat submitted that, with their core logistics models, both Deliveroo and
Uber Eats have typically focussed on cities and larger towns as they have
expanded, which provide a denser urban customer population and a large
accessible supply of drivers to scale their models quickly. Thus, Just Eat’s
competition with them has typically focussed on those areas.
5.116 However, Just Eat also submitted that it has witnessed both competitors
rapidly expanding their geographical spread into smaller areas. Just Eat told
us that, in June 2018, both Deliveroo and Uber Eats were present in
87
approximately 200 towns/cities. Since then they have more than doubled their
UK footprint, now both present in approximately 450 towns/cities. [].
5.117 Just Eat told us it predicts that the geographical expansion of its rivals will
continue to grow, []. Just Eat submitted that, in addition, Uber Eats and
Deliveroo’s investment in national advertising campaigns (eg Deliveroo on
Saturday Night Takeaway and the England Football Team and Uber Eats with
Love Island) only drive maximum returns if they have or are aiming for
national geographic coverage.
5.118 Similar to Deliveroo, Just Eat’s [].
5.119 Just Eat also makes some strategic decisions at a national level. For
example, internal documents provide details of national mass media TV
(eg X Factor, Ant & Dec’s Saturday Night Takeaway) and radio (eg Magic FM)
advertising campaigns that Just Eat have conducted.
• Uber Eats
5.120 Uber Eats submitted that the competitive landscape for online restaurant
platforms in the UK varies on a geographic basis. Uber Eats submitted that is
because:
(a) not all online restaurant platforms operate in all areas; and
(b) where Uber Eats operates, the presence and strengths of its main
competitors varies on a geographic basis.
5.121 Uber Eats submitted that this regional variation is due to the nature of
Delivered Prepared Food businesses, which are shaped by local factors such
as volume and quality of merchants
197
on the platform, number of customers
and number of couriers. Uber Eats submitted that these are, in turn, shaped
by factors such as time of entry to the market, marketing available delivery
methods, ability to recruit and retain talented employees (eg restaurant sales
teams) and quality of operations. It submitted that these factors mean that an
online restaurant platform’s offering varies in different parts of the UK.
5.122 Uber Eats also submitted that it understands that Just Eat’s logistics network
and delivery capabilities are currently limited to a few geographical areas in
the UK. Uber Eats therefore believes [].
197
[]
88
5.123 However, Uber Eats told us that, given Just Eat recently announced
partnerships with Greggs and McDonald’s, [].
5.124 In its internal competitive analysis documents, Uber Eats []. Uber Eats
notes that it has continued [].
• Our assessment
5.125 A number of factors point towards a national market in this case. In particular,
we note that:
(a) [] online restaurant platforms which are active in the UK have ambitions
to expand their geographic coverage nationwide.
(b) Deliveroo, Uber Eats and Just Eat all make some strategic decisions at a
national level, which will influence their effectiveness as a competitor
across multiple or all local areas. For example, they have invested in
national mass media advertising campaigns (eg Deliveroo on Saturday
Night Takeaway and the England Football Team; Uber Eats with Love
Island; and Just Eat with X Factor). Other aspects such as improvements
to logistic algorithms and app functionality, as well as terms of
membership to subscription services are determined centrally.
(c) Online restaurant platforms are competing for and signing up to exclusive
partnerships with nationwide restaurant chains such as McDonalds and
Greggs.
5.126 Nevertheless, there is evidence indicating that local variations in competitive
conditions also play a role, on both the supply and the demand sides:
(a) []
(b) Demand is inherently local, with consumers ordering from restaurants that
deliver to their address and restaurants listing on the Parties’ platforms in
order to tap into local consumer demand.
(c) []
(d) There are many geographical areas in which only one platform operates
(this is generally Just Eat, see Appendix C for geographical coverage
maps of market participants). However, as noted above, the internal
documents of [] indicate that they are looking to expand their services
nationwide.
(e) Geographical coverage maps (see Appendix C) of market participants (as
well as the submissions and evidence presented) highlight that the
89
intensity and closeness of current competition between suppliers of online
restaurant platforms is greatest in London and in a few other large cities.
Conclusion on geographic market definition
5.127 As set out above, we find that competition between the three online restaurant
platforms that are active in the UK has significant national aspects: [] have
ambitious plans to expand into additional UK locations. In addition, aspects of
their services which are determined centrally will influence their effectiveness
as a competitor across multiple or all local areas (eg advertising strategy,
improvements to logistic algorithms and app functionality, agreements with
large restaurant chains, terms of membership to subscription services). As
such, in this case we have primarily assessed competition on a national basis.
5.128 However, we have also considered local factors as part of our competitive
assessment, reflecting the fact that the demand for online restaurant
platforms’ services is local and that there are certain local variations in
competitive conditions.
Online convenience groceries
Introduction and overview
5.129 The starting point for our assessment of market definition is the overlapping
products of the Parties.
198
We have therefore considered the supply of online
convenience groceries (OCG) in the UK to be our focal product, that is,
groceries ordered online for delivery within a few hours.
5.130 This focal product includes Amazon Prime Now (for which the average
delivery time is []-hours) and Deliveroo (for which delivery is typically within
half an hour of ordering).
5.131 We have used the broad focal product of OCG to explore the broader
constraints that companies operating in this space may face, and to set out
where the market may be wider than this.
5.132 As set out in our competitive assessment, it appears that, on balance, while
the current offers of Amazon and Deliveroo are differentiated in range, price
and speed (reflecting their different business models) there may be some
overlap in the shopping missions which they serve. If so, there could be an
increasing competitive constraint between the two offers as Deliveroo’s offer
198
MAGs, paragraph 5.2.11.
90
becomes more widely established. We also consider whether the Parties’
OCG offers may become more similar as the market evolves.
5.133 The supply of OCG is a nascent market and suppliers are continuing to
experiment to determine the right way of addressing customers’ preferences
for greater convenience. As discussed in more detail in our competitive
assessment, it remains to be seen how the Coronavirus (COVID-19)
pandemic will affect the nascent OCG market in the medium- or longer-term.
As such, a static market definition has been given limited weight in this case,
as current substitution and product characteristics may not be reflected in the
future evolution of the market. The more important questions for this case are
how the market is likely to evolve, and the extent to which the Parties are
likely to constrain each other now and in the future. These questions are the
focus of our competitive assessment.
5.134 We evaluate the appropriate market definition by considering the extent to
which consumers and grocery suppliers can substitute between the Parties’
grocery offers and other grocery offers.
5.135 As explained below, we find that the relevant product market is OCG services
(ie groceries ordered online for delivery within a few hours). In particular:
(a) There is a distinction between online delivered groceries and OCG.
Slower online delivered grocery services are not a close substitute to
consumers for OCG and are, therefore, outside the market. However,
they are an out of market constraint.
(b) Bricks-and-mortar convenience stores do not compete closely with OCG
delivery, but they are an out of market constraint.
(c) Personal shoppers are in the market, but their activities in the UK are
extremely small.
(d) Grocery retailers supplying their own delivery are in the market at the
retail level. However, we note that grocery retailers have indicated that it
would be challenging for them to successfully develop their own cost-
effective OCG offer.
(e) Last-mile logistics specialists are in the market. A number of grocery
retailers are currently offering OCG services in partnership with a last-mile
logistics specialist.
(f) Traditional delivery providers (such as Royal Mail, Hermes and DPD) are
not well placed to offer ultrafast delivery on behalf of grocery retailers and
so are outside the market.
91
5.136 Our view on the relevant geographic market definition is that competition in
this market has important local factors (such as availability and potentially
pricing). However, several important parameters of competition are set
nationally. Given the nascent state of the market, we consider that it is not
appropriate to carry out a detailed assessment of local markets. We have
considered local factors in our competitive assessment.
5.137 The following analysis is largely based on evidence collected prior to the
Coronavirus (COVID-19) crisis. The Coronavirus (COVID-19) crisis has had a
substantial impact on the market for the supply of OCG, as discussed in
further detail in our competitive assessment. Based on our understanding of
these recent developments, we do not consider they have changed the nature
of demand for, or supply of, OCG to an extent that would require us to adopt a
different view of the relevant market.
Product market definition
5.138 We have considered the supply of OCG in the UK to be our focal point.
5.139 OCG services include:
(a) traditional and online grocery retailers who are either self-supplying last-
mile delivery (Amazon Prime Now) or outsourcing delivery to a last-mile
logistics specialist (eg Ocado Zoom, Sainsbury Chop Chop, Co-Op,
Morrisons); and
(b) partnerships between restaurant delivery platform operators and grocery
retailers, with the former providing app presence and/or delivery
(eg Deliveroo with Co-op, Uber Eats with Costcutter, Just Eat with Asda).
5.140 As set out in more detail in our competitive assessment, all of these market
participants are in the early stages of developing their OCG offer. Deliveroo is
in the early stages of developing its groceries business and supplies a limited
range of products to customers for delivery within half an hour. Amazon
supplies a wider range of groceries through Prime Now, available for delivery
within a few hours.
199
5.141 We have considered whether OCG providers face a competitive constraint
from the following:
199
We note that prior to 16 Match 2020, Amazon Prime Now also offered ‘click to deliver’ (CTD) and one-hour
scheduled delivery window options, but Amazon told us that this offer has now been withdrawn.
92
(a) Online delivered groceries;
200
(b) Bricks-and-mortar convenience stores and/or personal shoppers as
possible substitutes for supply for consumers; and
(c) Self-supply (like Amazon’s OCG business), last-mile logistics specialists
and/or traditional delivery providers as possible substitutes for grocery
retailers for grocery delivery.
5.142 Given the developing nature of this market, current competitive conditions
provide less of a guide to assessing the extent of future competitive
constraints. In this context, for the purposes of assessing future competitive
dynamics, we have considered carefully the Parties’ views, the evidence
contained in the Parties’ internal documents, as well as evidence from third
parties.
The Parties’ views
5.143 The Parties submitted that the segmentation of OCG as ‘groceries ordered
online for delivery within a few hours’, as something distinct from ‘online
delivered groceries’ is ‘an entirely artificial construct’.
The Parties submitted
that this segmentation places Deliveroo and Prime Now in the same market
by construction, ‘without evidence of current or future substitutability’, and that
this omits each of the Parties’ actual substitutes.
201
5.144 As discussed in more detail in our competitive assessment, the Parties
submitted that their propositions are fundamentally different and are not
substitutes for either consumers or grocery retailers. They are also based on
entirely different operating models.
202
(a) The Parties submitted that there is currently no interaction between the
Parties, and each has a distinct set of competitors (eg Just Eat and Uber
Eats for Deliveroo, and supermarkets for Amazon).
(b) The Parties submitted that Amazon and Deliveroo operate fundamentally
different services and do not compete for consumers, with Deliveroo’s
offer serving more urgent impulse purchases and Amazon’s offer being
closer to a weekly ‘big basket’ shop.
200
Ie groceries delivered in more than a few hours from ordering. This includes later same-day and next day
orders.
201
Deliveroo response to the revised provisional findings, 10 July 2020, paragraph 3.3.
202
Parties’ Initial Submission, 24 December 2019.
93
(c) The Parties submitted that they are based on entirely different operating
models and their different logistics networks prevent them from
competing.
203
Deliveroo’s point-to-point network achieves delivery speeds
of less than 30-minutes. It is limited (compared to Amazon) in the basket
sizes and product ranges it can offer. Amazon’s point-to-multipoint
network [].
(d) The Parties submitted that there is no competition for the supply of
logistics services between the Parties. Deliveroo supplies services to a
small network of convenience stores and petrol stations in connection with
a limited selection of convenience groceries, whereas Amazon provides
core shop grocery delivery services for Morrisons.
5.145 The Parties submitted that the offerings of online grocery retailers like Ocado
Zoom, Sainsbury’s Chop Chop, Shop Coop, Waitrose Rapid Delivery are also
differentiated from Amazon Prime Now and Deliveroo’s services. They do not
offer delivery within 20 to 30-minutes, as Deliveroo does, and they are also
differentiated from Amazon, targeting smaller baskets and one-hour delivery.
5.146 The Parties also submitted that, in light of their different operating models,
they are unlikely to compete in the future.
204
Our assessment
5.147 As outlined above we have considered the supply of OCG in the UK to be our
focal point.
5.148 We have considered the closeness of competition between the Parties in our
competitive assessment. Our view is that:
(a) Amazon and Deliveroo tend to be used for different shopping missions,
but there is some overlap in the shopping missions which they serve.
(b) There is scope for the Parties’ offers to evolve and become more similar
as the market develops as a result of (a) Amazon investing in faster
delivery and (b) Deliveroo expanding and improving its OCG offer.
(c) We note that Amazon and Deliveroo’s OCG offers are based on different
logistics networks (point-to-multi-point vs point-to-point). Each type of
network would present different challenges if the Parties wished to
expand their offers. However, as detailed in our competitive assessment,
203
Parties’ Initial Submission, 24 December 2019. Deliveroo reiterated this point in its response to our revised
provisional findings (Deliveroo response to the Revised Provisional Findings, 10 July 2020, paragraph 3.6).
204
Deliveroo response to the Revised Provisional Findings, 10 July 2020, paragraph 3.8.
94
we note that a range of other OCG providers are seeking ways to address
similar challenges in order to achieve delivery speeds of two-hours, one-
hour or faster.
(d) Deliveroo and Uber Eats developed their point-to-point networks for online
restaurant delivery, and at present their grocery offers over these
networks appear largely to cater for impulse shopping missions. However,
several grocery suppliers are using a point-to-point network for delivery
within one-hour,
205
but are not limited to impulse purchases (eg they offer
a relatively wide range, large baskets, and prices which are competitive
with in-store prices). We therefore disagree that there is a clear
separation between delivery via a point-to-point network and serving
shopping missions beyond impulse purchases.
Potential consumer-facing substitutes for OCG
Online delivered groceries
5.149 We have considered whether it is appropriate to consider OCG as a segment
that is distinct from the larger segment for online delivered groceries. For this,
we have explored whether consumers see later same-day or next day grocery
delivery services, as a substitute for groceries delivered within a few hours of
ordering. We have also considered whether grocery retailers and market
commentators see online convenience grocery offers as distinct from
traditional online grocery delivery services.
5.150 The CMA considered online grocery delivery in its April 2019 decision in
Sainsbury’s/Asda.
206
In Sainsbury’s/Asda, the CMA defined a product market
for ‘online delivered groceries’.
207
The CMA excluded services such as Prime
Now and the Co-op’s trial service, reflecting their more limited range and
different shopping missions, relative to the online groceries targeting big
basket shops.
208
5.151 In the past, grocery providers have developed online groceries delivery
operations in the UK which offer consumers the opportunity to complete a big
basket shopping mission, but which are not attractive for small basket or
urgent grocery missions. This is illustrated below in Figure 5.6, which was first
presented by the CMA in Sainsbury’s/Asda.
209
205
Using a last-mile logistics specialist, [].
206
Final Report, anticipated merger between J Sainsbury Plc and Asda Group Ltd, 25 April 2019.
207
Sainsbury’s/Asda, Final Report, 25 April 2019, paragraph 10.41.
208
Sainsbury’s/Asda, Final Report, 25 April 2019, footnote 518.
209
Sainsbury’s/Asda, Final Report, 25 April 2019, Figure 10.1.
95
5.152 As Figure 5.6 shows, the category ‘food/drink for right away’ has historically
accounted for almost no online shopping missions, and ‘smaller shop for next
2–3 days’ has historically accounted for relatively few online shopping
missions.
Figure 5.6: Online shopping missions
Source: CMA based on Tesco internal document.
• The Parties’ views
5.153 The Parties submitted that Amazon Prime Now (but not Deliveroo) competes
with the scheduled delivery offerings of supermarkets, and that:
(a) Items sold by Amazon Prime Now are priced [].
(b) Amazon faces a wide spectrum of competitors for core grocery shops.
These competitors include all of the main supermarkets, including Asda,
Iceland, Morrisons, Sainsbury’s, Tesco and Waitrose and (in the near
future) M&S (via Ocado), as well as Ocado.
(c) []. UK supermarkets are willing to work with Deliveroo, as well as with
Just Eat and Uber Eats to service shopping missions which they do not
otherwise service.
210
210
Parties’ Initial Submission, 24 December 2019, paragraphs 5.21 and 5.22.
96
5.154 The Parties submitted that Deliveroo does not compete with the scheduled
delivery offerings of supermarkets. Instead, retailers use Deliveroo in
conjunction with (and as a complement to) their other delivery models,
including their ‘big basket’ weekly shop delivery services and fast delivery
services including Amazon.
5.155 The Parties submitted that grocery retailers have well-developed point-to-
multipoint models and that, whilst they may face challenges in developing
quicker propositions, this appears to be a matter of investment which would
be made if there was enough demand for quicker services.
5.156 The Parties submitted that the difficulties that grocery retailers have indicated
they would face in successfully developing their own cost-effective quick
delivery services (such as those offered by Ocado Zoom, Waitrose Rapid or
Sainsbury’s Chop Chop) are no different to those Amazon would face.
• Evidence from the Parties’ internal documents
5.157 Amazon’s groceries plans distinguish [] ‘ultrafast competitors’, (as opposed
to ‘Next/Same Day Schedule Delivery’) []. Amazon’s internal documents
[].
5.158 As discussed in our competitive assessment, Amazon’s groceries plans
indicate that its Ultrafast Groceries service will supply [] products in the UK.
They show that it intends to compete [] (ie the supply of OCG, that is
groceries ordered for delivery within a few hours).
5.159 Amazon’s groceries plan from December 2019 indicates that while, in
principle, some grocery retailers’ existing non-convenience groceries delivery
services offer some degree of same-day delivery, [].
211
• Evidence from third parties
5.160 Market commentators have identified a gap between changing consumer
behaviour (increasingly shopping ‘little and often’) and online grocery
deliveries designed to serve the traditional big basket shop.
212
211
Amazon’s December 2019 grocery plan describes these as ‘[]’.
212
See, for example Online Grocery Retailing UK 2019 Mintel report, pages 20-21. The CMA identified a trend
toward shopping ‘little and often’ in its decision in Sainsbury’s/Asda noting that main-shop missions at
supermarkets declined from 49% to 42% between 2010 and 2018, and top-up and convenience store shopping
had increased (Sainsbury’s/Asda, Final Report, 25 April 2019, paragraphs 4.11–4.14).
97
5.161 In its March 2019 ‘Online Grocery Retailing UK’ report, Mintel discuss same-
day delivery services as a distinct group of propositions. It notes that ‘[t]he
potential benefit for the wider [online grocery retailing] market of this shift to
same-day is that it allows online grocery services to not just capture planned
big-basket supermarket-style shops, but also the more frequent smaller
basket shops that have become prevalent in grocery shopping behaviour
across the past decade’.
213
5.162 Grocery retailers have developed OCG offerings that are distinct from their
traditional online delivered grocery offerings (such as Ocado Zoom, Waitrose
Rapid, Sainsbury’s Chop Chop). Deliveries and logistics from these OCG
offerings are managed separately to their traditional online delivered grocery
orders (eg using a last-mile logistics specialist such as Stuart). However, we
note that this is not the case for Amazon Prime Now which offers both same
day and future slots under the same offerings and manages its delivery
windows together.
5.163 Ocado’s internal strategy documents discuss Ocado Zoom as being distinct
from its traditional online groceries offer. Ocado Zoom is described as [].
214
The paper includes the following slide (see Figure 5.7), [].
215
[]:
Figure 5.7: []
[]
Source: Ocado.
5.164 Tesco told us that instant delivery is currently a different shopping mission to
next-day delivery. However, in the long-term, as the instant delivery market
grows and becomes the norm for customers, shopping missions are likely to
converge, and customers are likely to shop for small and big baskets alike
through either instant or next day delivery. This is likely to be driven by the
proposition’s retailers’ offer in the market.
5.165 Uber Eats told us it currently offers an effective solution (as shown by the
demand response to its offer) for a somewhat urgent, last-minute,
complementary groceries shopping mission, [].
5.166 An Uber Eats internal document indicates that it views [] as competitors for
its online convenience grocery offer. A ‘convenience store strategy’
presentation refers to a range of competitors including [].
213
Online Grocery Retailing UK 2019 Mintel report.
214
[]
215
[]
98
5.167 Grocery retailers have also indicated that there is little scope for supply-side
substitution between the supply of online large basket groceries and the
supply of OCG (see paragraphs 5.204 to 5.208).
• Evidence from consumers
5.168 As discussed in more detail in our competitive assessment (paragraph 8.91),
we note that a majority of Amazon and Deliveroo customers responding to our
survey reported that they would not purchase groceries via an alternative
OCG provider if, hypothetically, their provider stopped offering the service.
216
18% of Amazon Prime Now customers reported that they would have ordered
groceries for next day or later from Amazon Prime Now. 18% of Amazon
Prime Now customers and 13% of Deliveroo customers reported that they
would have ordered groceries for next day or later from another supplier.
217
• Our assessment
5.169 We consider that it is appropriate to consider OCG services as being in a
distinct market from other online grocery delivery:
(a) There is little evidence of demand-side substitutability between the
traditional supply of online groceries and the supply of OCG. OCG
services (that deliver groceries within a few hours) are typically fulfilling a
different consumer need to online delivered grocery services for delivery
further in the future.
(b) Amazon’s Ultrafast Groceries service will supply a wide range of products
in the UK. We accept that traditional grocery delivery is likely to be a
substitute for part of Amazon’s grocery delivery offer, but do not view this
to be the case for its OCG offer (ie groceries delivered within a few hours
of ordering).
(c) There is also limited scope for supply-side substitution between the
supply of slower delivery of online large basket groceries and the supply
of OCG. Grocery retailers (with the exception of Amazon) have developed
OCG offerings that are distinct from their traditional online delivered
grocery offerings (such as Ocado Zoom, Waitrose Rapid, Sainsbury’s
Chop Chop). As outlined in more detail in paragraphs 5.204 to 5.208,
grocery retailers have told us that their existing delivery operations are not
suited to fulfilling OCG orders. Grocery retailers have told us that there
216
Accent and PJM Economics (April 2020), Final Report.
217
Accent and PJM Economics (April 2020), Final Report.
99
would be significant challenges associated with developing their own
OCG logistics and delivery capability.
(d) We note that OCG services have grown substantially both before and
following the Coronavirus (COVID-19) crisis, indicating that consumers
are willing to pay a premium (in the form of higher prices and/or delivery
charges) for these services. Clearly, the option of later online delivery will
place a limit on what consumers are willing to pay for OCG services.
However, in our view, competition between OCG services, rather than
with online delivery more generally, is likely to be the main driver of
pricing, speed and innovation as the market develops. This is supported
by internal documents which show OCG providers to be predominantly
focused on competition from other OCG offers.
5.170 Our conclusion is, therefore, that we consider traditional online delivery
grocery services to be out of the market. However, they are an out of market
constraint, which we consider, to the extent relevant, within our competitive
assessment.
Bricks-and-mortar convenience stores
5.171 We have considered whether offline sales of convenience groceries should be
included in the same product market definition.
5.172 The CMA considered online grocery delivery in its April 2019 decision in
Sainsbury’s/Asda.
218
The CMA defined a product market for ‘online delivered
groceries’.
219
The CMA distinguished between the convenience of in-store
and online groceries, noting that in-store provided ‘the ability to purchase
groceries immediately’ while online avoided ‘the need to make a specific trip
to a store’.
220
• The Parties’ views
5.173 The Parties submitted that shopping in bricks-and-mortar stores is still the
choice of the majority of consumers shopping for consumer products,
including groceries.
5.174 The Parties submitted that Amazon considers its competitors to be bricks-
and-mortar supermarkets (including their delivery capabilities) rather than on-
demand impulse/immediate delivery options such as Deliveroo’s offering.
218
Sainsbury’s/Asda, Final Report, 25 April 2019.
219
Sainsbury’s/Asda, Final Report, 25 April 2019, paragraph 10.41.
220
Sainsbury’s/Asda, Final Report, 25 April 2019, paragraph 10.24.
100
Items sold by Amazon Prime Now are priced [], and Amazon’s internal
documents [].
5.175 The Parties submitted that, in contrast, prices for products delivered using
Deliveroo’s service are typically at a very significant premium over
supermarket prices:
‘[]’.
5.176 The Parties submitted (prior to the Coronavirus (COVID-19) pandemic
221
) that
[].
• Evidence from third parties
5.177 A Co-op ‘Food Transformation’ steering group paper from 2019, which
presented results from a trial between Co-op and Deliveroo, stated that ‘[]’.
5.178 Just Eat submitted that ‘[]’.
• Evidence from consumers
5.179 As discussed in more detail in our competitive assessment (see
paragraph 8.91), we note that a majority of Amazon and Deliveroo customers
responding to our consumer survey reported that they would not purchase
groceries using an alternative OCG provider if, hypothetically, their provider
stopped offering the service, with the most common response being that they
would purchase in-store (31% of Amazon customers and 37% of Deliveroo
customers).
222
• Our assessment
5.180 Our conclusion is that bricks-and-mortar convenience stores are an out of
market constraint. Our basis for this position is set out below.
5.181 From a demand-side perspective, bricks-and-mortar stores, including
convenience stores, are currently the primary way consumers purchase
convenience groceries.
5.182 Consumers’ willingness to pay for delivery of groceries will be limited by their
willingness to switch to local convenience stores.
223
However, the supply of
221
The impact of the Coronavirus (COVID-19) pandemic on the market for OCGs is discussed in more detail in
our competitive assessment.
222
Accent and PJM Economics (April 2020), Final Report.
223
Ie the more willing/able you are to go to the shops instead, the less you will be willing to pay to have items
delivered.
101
OCG is attractive to customers who place a premium on convenience and are
therefore willing to pay for their groceries to be delivered to them within a
short time period.
5.183 In this context, we note that there is a price difference between OCG and
groceries available in bricks-and-mortar stores: in some cases, the products
themselves are more expensive, and customers pay a delivery fee (either for
the specific delivery or as part of a subscription fee for the delivery service). In
others, for example with Amazon Prime, the product prices are comparable to
those of supermarkets, but consumers are usually required to pay a delivery
fee if their order value is below a certain threshold.
224
Amazon Prime Now
customers are also required to be Amazon Prime members (which costs
£79 per year or £7.99 per month).
225
The fees charged by different suppliers
for their OCG offers are detailed in our competitive assessment.
5.184 We note that in-store prices are a comparison point for what suppliers can
charge for OCG, but OCG has grown while charging a premium on in-store, in
the form of delivery charges or higher unit prices. The scale of this premium
likely to be driven by competition between OCG providers rather than in-store
prices.
Personal shoppers
5.185 We have also considered whether personal shoppers should be included in
the market.
5.186 Personal shoppers (for example, Homerun) are services which operate a
delivery network through which customers can access groceries (and other
goods) at ultrafast speeds, without agreeing to a partnership with grocery
retailers. Their couriers purchase goods on behalf of the end-consumer by
walking into a store and picking the relevant products themselves, like any
other shopper.
• The Parties’ views
5.187 The Parties submitted that service providers like Home Run and Task Rabbit
compete with Prime Now in the UK. These entities allow customers to request
a personal shopper to select products in a store, including but not exclusively
groceries, and deliver their orders to them, as well as run other errands.
224
Amazon told us that it has temporarily removed Prime Now’s Free Shipping Threshold (FST) of £40 and
started delivering all Prime Now orders above the minimum order value of £15 with no delivery fee. Amazon told
us that it made this change [].
225
[].
102
5.188 The Parties submitted that they face competition from a range of personal
shoppers and other start-ups such as Homerun (a personal shopper),
Beelivery and Grocemania in the supply of online convenience groceries.
• Evidence from the Parties’ internal documents
5.189 Deliveroo’s internal documents briefly refer to personal shoppers, commenting
that they are ‘[]’ and/or ‘[]’, having raised relatively little capital.
• Evidence from third parties
5.190 Questionnaire responses from personal shoppers such as Grocemania and
Beelivery indicate that they view themselves as being active in the market for
on-demand convenience groceries.
5.191 The activities of such companies in the UK are currently small. One of the
largest of these companies, Homerun, generated revenue of under £1 million
in 2018.
226
Another of these companies ([]) had expected total revenue in
2019 of []. [] had total revenue in 2019 of £[] million, with []% of this
delivered through a logistic partner’s delivery network.
5.192 Co-op, who has a partnership with Deliveroo amongst others, told us that it
has launched a partnership with personal shopper service Buymie and [],
offering same day and on demand propositions. These platforms match users
with a local personal shopper who will pick and purchase from Co-op stores
and then deliver shopping within a short window of time to users.
• Our assessment
5.193 Our view is that personal shoppers are in the relevant product market. In
principle, personal shoppers have the potential to offer a wide range of
selection from multiple grocery retailers and compete in the supply of OCG.
Grocery retailers are trialling their services for online convenience grocery
delivery.
5.194 However, the evidence we have indicates that the activities of such
companies in the UK are currently extremely small.
5.195 We also note that this type of business is likely to be at a cost disadvantage to
providers such as Deliveroo as they do not typically receive a commission
from grocery retailers for providing this service, which could inhibit their ability
226
See: Seedrs Limited website.
103
to expand and provide a competitive offering to the Parties. However, we note
that Co-op told us that it was exploring offering [].
Potential substitutes for grocery delivery
5.196 We have defined the retail market to include all OCG offers, including those
where the grocery provider operates its own delivery service, provides the
offer in partnership with an online restaurant delivery platform, or contracts
with a third-party delivery provider. In the following, we consider the delivery
and consumer interface (eg app or website) alternatives available to grocery
providers.
5.197 As set out in our competitive assessment, to reach consumers with an OCG
offering, grocery retailers need (i) fulfilment centres, (ii) an ecommerce
website or app, and (iii) a delivery network capable of picking and delivering
groceries at ultrafast speeds (ie for delivery within a few hours).
5.198 Grocery retailers’ existing stores can act as fulfilment centres, with existing
staff picking goods. However, grocery retailers also need the logistics for rapid
delivery, which in the case of Morrisons is provided by Amazon.
Evidence from third parties
5.199 Grocery retailers told us that they can produce an ecommerce website
themselves, or they can partner with a supplier which already operates an
ecommerce website or app.
227
Several grocery retailers told us that they
preferred to use their own ecommerce website, as this gave them greater
control over the customer relationship,
228
although some grocery retailers
noted that customer acquisition costs for propositions on their own websites
could be more expensive than customer acquisition through online food
platforms, and that using an alternative platform may give them access to
customers they would not otherwise reach.
229
5.200 Grocery retailers told us that a key challenge to offering OCG is arranging
delivery.
230
For example, one retailer told us that ‘[f]inding a solution to the
cost of providing an ultrafast delivery service is the biggest economical
challenge in the online delivery space. The largest costs come from the fixed
227
For example, Co-op told us that it partnered with [] to provide the Co-op with an on-demand grocery
delivery platform so that the Co-op could fulfil its own e-commerce orders and deliveries.
228
[] notes that a key advantage to self-supply would be greater ownership of the customer relationship from
point of sale to fulfilment.
229
[] notes that the key advantage to using a food delivery marketplace would be access to a large customer
base who are already engaged with on-demand services and an existing on-demand delivery network.
230
For example, [] and [] (which said it ‘considers that the big challenge for all players is the logistics of on-
demand grocery delivery’).
104
costs such as getting vans on the road or employing people to sit and wait for
orders to be placed’.
5.201 We have therefore considered whether each of the below methods of
arranging groceries delivery should be included in the supply of OCG.
Self-supply
5.202 As set out in more detail in our competitive assessment, a number of
traditional and online grocery retailers are offering an OCG service through
outsourcing delivery to a last-mile logistics specialist, or to a restaurant
delivery provider (eg Co-Op with Deliveroo, Costcutter with Uber Eats, Asda
with Just Eat).
5.203 We first considered whether these and other traditional and online grocery
retailers could use their existing delivery operations to fulfil online
convenience grocery orders.
• Evidence from third parties
5.204 Third parties have indicated that there is little scope for supply-side
substitution between the supply of online large basket groceries for delivery
later same-day or further in the future and the supply of OCG:
(a) The majority of grocery retailers responding to our third-party
questionnaire told us that they have considered fulfilling the delivery and
logistics for their online convenience grocery service internally, but there
are challenges associated with developing their own on-demand logistics
network.
(b) Grocery retailers have told us that their existing delivery operations are
not suited to fulfilling urgent grocery orders. They told us that to offer this
they would need significant upfront capital expenditure, to expand
resources and operational and technological capabilities, and risk
potential store disruption.
231
Instead, grocery retailers have trialled
alternative models of groceries delivery using third-party delivery couriers.
Examples include Tesco Now, Sainsbury’s Chop Chop, Waitrose Rapid
and Ocado Zoom (orders are mainly fulfilled by Stuart).
(c) One grocery retailer submitted that it had considered fulfilling the delivery
logistics for its OCG service internally. However, following analysis, found
231
[] told us they need partners for the technology and couriers, therefore their existing operations are not
suited to fulfil urgent grocery orders. In grocery retailers’ responses to the phase 1 questionnaires, there were a
lot of references to the ‘cost and complexity’ of trying to self-supply.
105
that it was not economically viable to maintain an in-house fleet at the
requisite scale, and that having dedicated bikes or vans for on-demand
delivery was too expensive. It will continue to consider whether there is
any economic value in integrating stand-alone on-demand delivery
services into its main website and using its existing fleet to render these
delivery services, but its current viewpoint was that this presents several
significant challenges from both a technical and operational perspective.
(d) Another grocery retailer submitted that it had not considered fulfilling the
delivery logistics for its OCG service internally due to the amount of
resource required to create the logistics chain.
(e) Another grocery retailer submitted that fulfilling the delivery logistics for its
OCG service internally is not its preferred route due to the cost and
complexity associated with creating and maintaining its own fleet of
vehicles and drivers.
(f) One grocery retailer told us that [].
(g) Grocery retailers told us that a key challenge in offering OCG is achieving
sufficient scale to justify the costs of a courier delivery operation. For
example, one of these grocery retailers told us that the costs and
resource of ‘maintaining a fleet that is able to support an on-demand
proposition make this model operationally complex and expensive’.
5.205 One last-mile logistics specialist also told us about the importance of
achieving sufficient scale and scope to make the economics work. It told us
that ‘[]’.
5.206 A number of grocery retailers did however tell us that they see self-supply as
their best potential route to market for on-demand groceries. Grocery retailers
highlighted a number of benefits of self-supply, including that this would allow
them to retain control of the customer relationship and supply chain, and
enable them to avoid paying commission fees.
5.207 One grocery retailer also submitted that it has considered fulfilling the delivery
logistics for its OCG service internally, potentially using its existing fleet. It
submitted that it has an existing framework in place due to its existing (offline)
home delivery business, which delivers groceries the same day for customers
that have physically shopped in some of its stores.
• Our assessment
5.208 As discussed, we have included all OCG services at a retail level. The ability
of grocers to provide an OCG service depends on their ability to create or
106
access an on-demand delivery network (and consumer interface). Creating
such a service is challenging, but some grocers are doing this. The main third-
party alternatives for grocers to work with to offer an OCG service are
contracting with Stuart, or partnering with Deliveroo, Uber Eats, or Just Eat.
Last-mile logistics specialists
5.209 We have also considered whether grocery retailers could use last-mile
logistics specialists to fulfil online convenience grocery orders.
• The Parties’ views
5.210 The Parties submitted that last-mile logistics specialists provide grocers with a
range of credible providers, including Stuart and City Sprint.
5.211 The Parties submitted that ‘Stuart (used by Sainsbury’s Chop Chop and
Ocado Zoom), CitySprint (used by Waitrose Rapid
232
), and eCargo bikes
(used by Co-op “Quick Shop”) already provide logistics to grocers as part of
their experimental offerings for smaller baskets delivered more quickly, and
others could’.
5.212 The Parties also submitted that the fact that some last-mile logistics
specialists do not offer a sales channel is unlikely to be a significant issue.
The supermarkets have powerful brands and can easily develop, or outsource
the development of, a dedicated app. This also avoids their products being
listed alongside competitors’ products and allows them to retain control of the
customer relationship. Moreover, this has been the approach for all of the
experimental offerings between grocery retailers and last-mile logistics
specialists to date.
• Evidence from third parties
5.213 Third parties told us that one last-mile logistics specialist, Stuart, has
demonstrated strong ultrafast delivery capabilities, while also mentioning that
other last-mile logistics specialists were less capable of delivering chilled or
hot food on-demand. Third parties also told us that some last-mile delivery
companies that entered the UK (such as Quiqup and Jinn) were not
successful and as such were not options.
233
232
We note that Waitrose Rapid now use Stuart (as of 5 June 2020).
233
For example, [] told us that it used On the Dot but were not happy with it and thought Stuart was probably
better, and everyone else who are already using Stuart ([]) said they thought it was their best available
provider.
107
5.214 Stuart submitted that ‘it does not view []. It considers that it only competes
to a limited extent with []’.
5.215 Stuart provides OCG delivery services to Co-op, Sainsbury’s and Ocado. [].
Stuart submitted that [].
• Our assessment
5.216 Our conclusion is that last-mile logistics specialists are in the market. A
number of grocery retailers are currently offering OCG services in partnership
with a last-mile logistics specialist. They are a competitive constraint and
should be included in the same product market.
5.217 A limited number of courier services specialise in last-mile delivery to
customers, at ultrafast speeds. Several grocery retailers continue to trial
ultrafast delivery with these last-mile logistics providers. More detail on these
trails is provided in our competitive assessment.
Traditional delivery providers
5.218 We also considered whether grocery retailers could use traditional delivery
providers (such as Royal Mail, UPS and DPD) to fulfil online convenience
grocery orders.
• The Parties’ views
5.219 The Parties submitted that grocery retailers have several delivery options,
including established logistics companies such as FedEx and Hermes.
5.220 The Parties submitted that ‘these providers are well placed [as possible
substitutes of supply for grocery retailers]. DPD owns Stuart, who the CMA
considers are in the OCG frame of reference. In addition, they have a deep
understanding of logistics, huge customer bases, strong brands and point-to-
multipoint capabilities’.
5.221 The Parties submitted that ‘Amazon faces the same challenges that would be
faced by traditional delivery providers to developing cost-effective ultrafast
delivery services’.
• Evidence from third parties
5.222 Traditional couriers told us that their existing delivery operations are not well
placed to offer ultrafast delivery.
108
5.223 Grocery retailers confirmed that they considered using traditional delivery
providers for convenience groceries, however they also told the CMA that
they considered traditional couriers inappropriate for offering same day or
shorter delivery of food. For example, one grocery retailer told us that
traditional couriers such as DPD are not set-up for same-day delivery and the
costs are prohibitive.
• Our assessment
5.224 Our view is that traditional delivery providers (such as Royal Mail, Hermes,
UPS and DPD) are not a substitute of supply for grocery retailers and are,
therefore, outside the relevant product market.
• Conclusion on product market
5.225 Our view is therefore that the relevant product market in which to assess the
effect of the Proposed Merger is the market for the supply of OCG, that is
groceries ordered online for delivery within a few hours. This includes
personal shoppers as possible substitutes for supply for consumers; and last-
mile logistics specialists as possible substitutes of supply for grocery retailers.
Self-supply by grocery retailers supplying their own delivery is also in the
market at the retail level. However, we note that some grocery retailers have
indicated that it would be challenging for them to develop their own cost-
effective OCG offer.
Geographic market definition
5.226 In the following section we have explored whether evidence indicates the
market for OCG is local or national in scope.
• The Parties’ views
5.227 The Parties submitted that the CMA assessed the geographic scope of online
delivered groceries in Sainsbury’s/Asda
234
and found this to be narrower than
UK-wide from both the demand and supply side. The Parties submitted that
from the consumer perspective, the geographic market is local in character
while from the supplier perspective the geographic market is the delivery area
served by each supply point (eg normal stores with store-pickers or dedicated
online delivery stores).
234
Sainsbury’s/Asda, Final Report, 25 April 2019.
109
5.228 The Parties submitted that [].
5.229 The Parties submitted that where Deliveroo is already active with its
restaurants it may be able to carry some bolt-on ‘grocery’ items.
• Evidence from the Parties’ internal documents
5.230 The Parties’ internal documents indicate that they are planning or considering
expanding their geographic reach. For example:
(a) [].
(b) Deliveroo’s internal documents also indicate that it has ambitions to
expand its []. We note this would give them an opportunity to offer
wider delivery of OCG services.
(c) []:
‘[]’.
• Evidence from third parties
5.231 Details of third parties’ entry and expansion plans are presented in our
competitive assessment. In summary, we find that most current and potential
market participants have expressed an interest or intention to enter or expand
their presence in the OCG market, and interest in growing in the market was
shared across different types of market participant – online grocery, online
restaurant marketplace, convenience stores and traditional grocers.
5.232 As discussed in our competitive assessment, a number of providers have
advanced their current and planned OCG services in response to the
Coronavirus (COVID-19) crisis. However, it remains to be seen how it will
affect the nascent OCG market in the medium- or longer-term.
• Our assessment
5.233 Competition in this market has an important local element. In particular, we
note that:
(a) The Parties compete locally to supply OCG to consumers. They also
leverage their existing local grocery supply chain and delivery capabilities
to provide ultrafast delivery capabilities to grocery retailers.
(b) For a market participant to compete effectively in OCG they need to have
a grocery supply chain (ie delivery of groceries to a chain of stores and/or
110
fulfilment centres). There will be a limited delivery range around those
stores and/or fulfilment centres.
(c) Last-mile delivery is something that needs to be achieved and made to be
economically viable in each location where OCG services are offered.
Maintaining a last-mile delivery network is expensive and relies on
achieving a sufficient scale of orders at a local level to justify the costs.
Achieving the required scale will be easier in more densely populated
areas.
5.234 However, a number of factors point towards a national market in this case. In
particular, we note that:
(a) Several important parameters of competition are set nationally, for
example, [].
(b) Improvements to an online restaurant platform’s core business (eg more
users and traffic on its app, improved logistics algorithms) will tend to
make it a more effective competitor in OCG services, so to the extent that
online restaurant platform competition is national, this will also be true of
OCG services.
(c) As discussed in our competitive assessment, we expect that traditional
grocery retailers will increasingly compete in OCG provision. These
traditional grocery retailers have a national presence.
(d) Partnerships between online restaurants platforms and grocery retailers to
provide OCG services typically cover multiple local areas.
5.235 Furthermore, the supply of OCG is a nascent market, in which providers are
still trialling and experimenting with their offering or are only beginning to roll-
out their services more broadly. The Parties themselves are embarking on
programmes to expand their geographic reach.
• Conclusion on geographic market definition
5.236 In conclusion, our view is that competition in this market has important local
aspects (such as availability and potentially pricing). However, several
important parameters of competition are predominantly set nationally. Given
the predominance of national competition, and the nascent state of the
market, we find that it is not appropriate to carry out a detailed assessment of
local markets. We have considered local factors where relevant in our
competitive assessment.
111
6. Counterfactual
Legal framework
6.1 The counterfactual is an analytical tool used to help answer the question of
whether a merger has or may be expected to result in an SLC. It does this by
providing the basis for a comparison of the competitive situation on the market
with the merger against the most likely future competitive situation on the
market absent the merger.
235
The latter is the counterfactual.
236
6.2 We may examine several possible scenarios to determine the appropriate
counterfactual, one of which may be the continuation of the pre-merger
situation (ie the prevailing conditions of competition). An example of a
situation where the CMA may select a counterfactual different from the
prevailing conditions of competition is where the target is likely to exit the
market absent the transaction under review. Another scenario in which the
CMA may consider an alternative counterfactual to the prevailing conditions of
competition is where one of the merging parties would have entered or
materially expanded its presence in a market absent the transaction.
6.3 We seek to avoid importing into the assessment of the appropriate
counterfactual any spurious claims to accurate prediction or foresight. Given
that the counterfactual incorporates only those elements of scenarios that are
foreseeable, it will not in general be necessary to make finely balanced
judgements about what is and what is not included in the counterfactual.
237
In reaching a view on the appropriate counterfactual, we must determine what
future developments we foresee arising absent the merger based on the
evidence available to us.
6.4 Events which occur during the CMA’s review of a transaction, but which are
not a result of the merger, can be incorporated into the counterfactual.
238
Where future events or circumstances are not certain or foreseeable enough
to include in the counterfactual, the analysis of such events can take place in
the assessment of competitive effects.
239
235
MAGs, paragraphs 4.3.1 and 4.3.6.
236
MAGs, paragraph 4.3.1.
237
MAGs, paragraphs 4.3.2 and 4.3.6.
238
See, in this regard, British Sky Broadcasting Group plc v Competition Commission (CC) (Cases 1095/4/8/08
and 1096/4/8/08) [2008] CAT 25, paragraph 138.
239
MAGs, paragraph 4.3.2.
112
6.5 Where there is more than one possible alternative scenario, the CMA will
select the counterfactual it considers would be the most likely scenario to
have arisen absent the merger.
240
6.6 Against this framework and in light of the Parties’ and third parties’
submissions and the evidence we observed, we considered the likely future
situations of both Amazon and Deliveroo in the absence of the Transaction.
Our focus was on two key questions: (a) whether Deliveroo would exit the
market in the absence of the Transaction; and (b) whether Amazon would re-
enter the market for online restaurant platforms.
Counterfactual for online convenience groceries
6.7 Insofar as the market for online convenience groceries (OCG) is concerned,
we considered the developments and expansion plans of the Parties and
whether it was appropriate to incorporate elements of these plans into an
alternative counterfactual. During its phase 1 investigation, the CMA assessed
possible changes in the market, including the expansion plans of the Parties
and their competitors, in its competitive assessment rather than as part of the
counterfactual.
6.8 For our phase 2 investigation we have adopted a similar approach to
assessing future competition in OCG. The Parties have not expressed
concerns with this approach or made submissions on what they consider the
appropriate counterfactual to be. We have sought to conduct a forward-
looking assessment concentrating not only on what the Parties and their
competitors currently offer but on what developments are planned or expected
in the future and the implications for competition. The online convenience
grocery market is nascent and many market participants are at an early stage
of trialling their offerings, or gradually rolling them out across geographical
areas. Views vary considerably as to how the market will develop, both
among providers and industry analysts. This uncertainty is further increased
by the recent changes brought about by Coronavirus (COVID-19), which has
impacted and will continue to impact this market. Because in our view future
OCG market developments are not sufficiently foreseeable to include in a
counterfactual, we have considered the likely impact of the Transaction based
on an in-the-round assessment of how competition may develop in future.
240
MAGs, paragraph 4.3.6.
113
Deliveroo counterfactual
Overview of exiting firm counterfactual
6.9 As noted above, one situation where the CMA may consider a counterfactual
that is different from the prevailing conditions of competition (the pre-merger
situation) is the ‘exiting firm scenario’. When considering an ‘exiting firm
scenario’ the CMA examines whether the firm would have left the market
absent the transaction and the impact of allowing the transaction to proceed.
The CMA applies a three-limb test when making this assessment,
considering:
(a) Whether the firm would have exited (through failure or otherwise) absent
the transaction.
(b) Whether there would have been an alternative purchaser for the firm or its
assets.
(c) What the impact of exit would be, and how this would compare to the
impact of the acquisition.
6.10 In considering whether Deliveroo should be considered to be an exiting firm
we have sought to understand the impact of the Coronavirus (COVID-19)
crisis on its business by considering its actual financial position. At the same
time, we have sought to abstract from all Transaction-specific elements of its
actual financial position in order to describe the relevant counterfactual.
Summary of April Provisional Findings analysis
6.11 In the April Provisional Findings,
241
we noted that, as a loss-making business,
Deliveroo has historically been reliant on external funding in order to continue
trading. We found that, at the time the Series G funding round (that included
the Amazon funding) was agreed, Deliveroo had alternative potential funding
options available to it, and in the absence of the Transaction Deliveroo would
have succeeded in raising some alternative funds. We also found that this
amount would not be likely to exceed the $[] million received to date from
the Series G funding round.
6.12 Our view at that time was that, in the counterfactual, Deliveroo would likely be
close to a position before the start of the Coronavirus (COVID-19) crisis where
it would need to raise additional funding and that it was likely that an
additional fundraising round would be required at some point in 2020.
241
April Provisional Findings, 16 April 2020.
114
However, our view was that absent the Coronavirus (COVID-19) crisis it is
likely that it would both have sufficient time to do so and that such funding
would be available. This is consistent with the Parties’ submission in the
Merger Notice that, absent the Transaction, Deliveroo ‘would have continued
to compete as it currently does, including by seeking suitable investment to
drive its expansion and innovation’. On this basis, Deliveroo would not have
been considered a failing firm.
Initial impact of Coronavirus (COVID-19) on Deliveroo
6.13 At the time of the April Provisional Findings, the CMA found that the
Coronavirus (COVID-19) crisis had had a severe impact on Deliveroo’s
business.
6.14 Deliveroo told us it had seen a significant decline in orders, with orders in the
last three weeks of March 2020 down []% in the UK and []% in its []
(France). This impact, Deliveroo stated, was due to a combination of
restaurant closures, perceptions of the safety of takeaway food and the
economic impact of the crisis on disposable incomes.
6.15 This, in turn, had a significant impact on Deliveroo’s profitability and cash
position. Based on its then current predictions of the ongoing impact of the
Coronavirus (COVID-19) crisis, Deliveroo forecast that it would only have
£[] in cash by the end of June 2020, despite the impact of significant cost
savings and including £[] in deferred tax and creditor liabilities (that it would
have to repay in the short-term). As a result of Deliveroo’s deteriorating cash
position, Deliveroo’s directors were advised by [] and their legal advisors
[] that if they did not have a reasonable expectation of receiving additional
funds before Deliveroo ran out of cash in early Q3 then they would be
required to declare Deliveroo insolvent [] as a result of its cash flow
problems.
6.16 Whilst Deliveroo was able to defer some taxes it told us that it was not eligible
for any of the corporate government support schemes available at the time of
the April Provisional Findings. It also told us that the furlough scheme would
only have a limited impact as [].
6.17 Deliveroo made several submissions in relation to how this position should be
taken into account within the CMA’s investigation:
(a) On 20 March 2020, Deliveroo submitted a letter from its board of directors
explaining that, as a result of the Coronavirus (COVID-19) crisis they were
likely to need to start insolvency proceedings []. Deliveroo’s letter
stated that it was ‘critical’ for Deliveroo that the CMA delivered its
115
provisional findings as soon as possible, and that it indicated in those
findings it was minded to clear the Transaction.
(b) On 26 March 2020, Deliveroo submitted that the impact of the crisis
meant that the relevant counterfactual was one in which Deliveroo would
exit the market. This was supported by various trading and financial
updates, including professional advice from third-party legal, financial and
insolvency advisors.
(c) Deliveroo provided updates to the CMA, including material from its
advisors and responses to queries on these submissions. Shortly before
publication of our April Provisional Findings, we also received updated
information from Deliveroo and confirmation that at that date there had
been no material recovery in the business.
242
Impact of Coronavirus (COVID-19) on the April Provisional Findings
counterfactual
6.18 We considered Deliveroo’s evidence in the context of the relevant
counterfactual (ie the situation that would have existed absent the
Transaction). In particular, in order to exclude the possibility that the
Transaction itself could be a contributing factor to Deliveroo’s exit, we
considered what alternative funding was available to Deliveroo at the time it
agreed the Transaction and what position it would likely be in had it accepted
that alternative funding. We determined that, absent the Transaction,
Deliveroo would have successfully raised some alternative funding at the time
it agreed the Series G round and, whilst the total amount was uncertain, that
this would not exceed the $[] million it had received to date from the Series
G round. We therefore provisionally concluded that this alternative funding
would have resulted in Deliveroo needing to raise additional funding in early
2020. We also provisionally concluded that the impact of Coronavirus
(COVID-19) we had observed on orders, restaurant availability and, in turn,
cash flow, would have been similar in the counterfactual. Accordingly, we took
the view that Deliveroo’s financial position at the time of the April Provisional
Findings and, in particular, its urgent need to raise funds could not be
attributed to its decision to accept investment from Amazon rather than from
242
Deliveroo has argued that the CMA misrepresented Deliveroo’s submission, as the key point of that
submission was that Deliveroo continued to face imminent insolvency. The paragraphs of the RFI response cited
by the CMA state the following: []. We do not consider that its assessment misrepresents these statements
from Deliveroo, nor that, as Deliveroo appeared to suggest in its response to the Revised Provisional Findings,
we should focus on statements made in the executive summary of Deliveroo’s response rather than on the
detailed responses to particular questions.
116
an alternative investor that might have been less likely to raise competition
concerns.
6.19 In respect of the ‘Limb 1’ analysis (whether the firm would exit in the absence
of the transaction), our provisional view was therefore that Deliveroo would
have been likely to exit the market as a result of the Coronavirus (COVID-19)
crisis without additional funds.
6.20 We then considered ‘Limb 2’ of the exiting firm scenario (whether there would
be an alternative purchaser for the business or assets (or in this case an
alternative investor in Deliveroo)). We considered the evidence that Deliveroo
had provided of the efforts it had made to secure urgent alternative
investment but, as for Limb 1, sought to abstract from any Transaction-
specific issues Deliveroo might have encountered in order to determine what
the position would have been in the counterfactual.
6.21 We found that as a result of the sudden and severe deterioration in
Deliveroo’s finances, it would need to obtain investment both in an extremely
short time frame and at a time when funding markets were severely impacted
by the Coronavirus (COVID-19) crisis. Our view was that a combination of
these factors meant that Deliveroo would be unlikely to be able to raise the
funding required within this timeframe. This was supported by submissions,
and responses to statutory information requests, from Deliveroo’s financial
advisors [].
6.22 We did not consider that the issues outlined above would differ materially if, in
the counterfactual, Deliveroo had pursued one of the alternative funding
options available in 2019 and therefore provisionally concluded that the most
likely counterfactual for Deliveroo absent the Transaction would have been its
exit from the market.
6.23 This provisional conclusion was driven primarily by the following combination
of exceptional factors:
(a) the particular structure of Deliveroo’s business and its need for regular
external funding to continue to grow its business and compete;
(b) the effects of the Coronavirus (COVID-19) crisis on both Deliveroo’s
financial performance and its ability to raise additional funds; and
(c) the particular point in Deliveroo’s funding cycle that Coronavirus (COVID-
19) occurred when, regardless of the transaction pursued in May 2019, it
would have had limited remaining funds available.
117
6.24 For the purpose of assessing ‘Limb 3’ of the exiting firm scenario (the impact
of Deliveroo’s exit on competition), we assumed that Amazon would have re-
entered the market absent the Transaction and considered whether this would
offset the loss of competition that would result from Deliveroo’s exit. We noted
that the exit of Deliveroo ‘would result in significantly weaker competition over
an extended period of time, and this would be the case even if Amazon
ultimately re-entered the market successfully’.
243
In respect of the online
convenience grocery market, we noted that ‘the loss of Deliveroo from the
market […] would likely reduce competition in the market immediately and
would reduce pressure on all suppliers to introduce or enhance point-to-point
delivery networks in future’.
244
Our provisional view was therefore that ‘Limb 3’
of the exiting firm scenario would also be satisfied.
Developments since the April Provisional Findings
6.25 Since the April Provisional Findings, the situation has continued to evolve and
there have been a number of key developments including:
(a) There has been a recovery in order volumes in the online restaurants
platform market. In addition, demand for takeaway food has continued to
increase as customers switched to dining in (from dining out).
(b) There has been sustained demand for online grocery delivery services,
with online restaurant delivery platforms expanding their operations in this
area.
245
(c) In the UK, lockdown measures have gradually eased with the UK
government announcing an initial easing of measures on 13 May 2020
246
with devolved administrations announcing easing of restrictions in the
following weeks. This has led to the re-opening of many restaurants,
including larger chains such as McDonalds
247
and Pret A Manger.
248
6.26 The CMA has also received some evidence that funding markets (which were
severely impacted by the crisis) have started to recover. For example, third
parties highlighted to the CMA that online restaurant platforms such as Just
Eat Takeaway.com and Wolt Group, (a Finnish restaurant delivery platform)
have announced significant external equity and debt funding (on 23 April 2020
243
April Provisional Findings, 16 April 2020, paragraph 4.80.
244
April Provisional Findings, 16 April 2020, paragraph 4.88.
245
Just Eat Takeaway.com response to the April Provisional Findings, 11 May 2020, paragraph 28.
246
See BBC, 13 May 2020, Coronavirus: Some return to work as lockdown eases slightly in England.
247
See McDonalds Newsroom, 12 May 2020, Reopening update for McDonald's UK & Ireland.
248
See Independent, 28 May 2020, Pret a Manger is reopening more than 200 sites across the UK from 1 June.
118
and 15 May 2020)
249
since the announcement of the provisional findings on
17 April. DoorDash also recently announced (on 18 June 2020) significant
external equity funding.
250
A recovery in the share prices of most listed online
delivery platforms in the period from the middle of March 2020 to the middle of
June 2020 in excess of the performance of the relevant market averages.
further supports the position that the businesses in this sector are now in a
stronger position to secure additional funding where necessary.
Figure 6.1: Movement in share prices of listed online delivery platforms in the period from the
middle of March 2020 to the middle of June 2020
Source: CMA analysis.
Notes: Red line indicates the WHO’s announcement of Coronavirus (COVID-19) as Global Pandemic on 11 March 2020.
Impact of developments on Deliveroo
6.27 Since the publication of the April Provisional Findings, we requested, and
have been provided with updated financial information from Deliveroo. This
shows that Deliveroo’s actual performance in April, May and June 2020 was
249
Just Eat announced it had raised €700 million worth of equity (€400 million) and convertible bonds
(€300 million) on 23 April 2020. See Takeaway.com Press releases, Just Eat Takeaway.com successfully raises
EUR 700 million through an accelerated bookbuild offering of new shares and convertible bonds and Bloomberg,
15 May 2020, Wolt Draws Goldman Investment in $108 Million Funding Round.
250
DoorDash announced it had raised $400 million in equity on 18 June 2020. See DoorDash.com blog ‘Investing
in the future of local commerce. DoorDash raises Additional Capital from New Investors’.
119
significantly better than had been forecast in March 2020. Deliveroo’s actual
‘EBITDA-capex’
251
[] for the period was £[] compared to £[] forecast at
that time. This is driven by both a continued increase in orders
252
(order
volumes for April while still below Deliveroo’s initial budget, were []% higher
than had been forecast in March; May and June order volumes exceeded
those forecast in Deliveroo’s pre-crisis budget in February by []% and []%
respectively) and a [] in April.
253
Deliveroo has told us that the [] was due
to a shift in mix of orders to [] and an increase in average order size due to
more families and whole households ordering. Online convenience grocery
deliveries have also increased significantly but remain a relatively small part
of Deliveroo’s business overall.
6.28 This has significantly improved the position from the prior forecast, resulting in
Deliveroo’s forecast losses now being £[] less, for the three months from
April 2020 to June 2020, than in its forecast at the time of the April Provisional
Findings. This in turn, combined with a £[] improvement in working capital,
means that Deliveroo had £[] of cash in June 2020
254
as opposed to the
£[] it had previously forecast.
6.29 Deliveroo’s revised forecast for the year to 31 December 2020 now shows an
EBITDA-capex [], a substantial improvement compared to an EBITDA-
capex [] predicted in the updated budget prepared in February 2020. This
has been achieved due to a combination of improved performance since the
end of March 2020, which accelerated through April 2020, and cost savings
compared to both budgets. Deliveroo’s current cashflow forecasts show it now
has an improved [] prior to the Coronavirus (COVID-19) crisis, albeit that,
as Deliveroo has noted, the current crisis also results in a higher degree of
uncertainty compared to pre-Coronavirus (COVID-19) forecasts.
6.30 We note that, given this longer cash runway, restructuring options that were
previously not feasible to Deliveroo, as they would be cash negative in the
short-term, such as rationalisation of its international operations, may now be
possible and could extend this runway further.
6.31 In summary, consistent with the submissions received by the CMA from third
parties, Deliveroo experienced a material detrimental financial impact at the
251
EBITDA: Earnings before interest, taxes, depreciation and amortization. Capex: capital expenditure.
252
Orders increased by [] in April actuals vs previous forecast at the start of the crisis. May and June orders
forecast to by [] and [] respectively.
253
Net revenue per order increased by £[] per order in April actuals vs previous forecast at the start of the
crisis.
254
This figure excludes liabilities that Deliveroo has been able to defer (such as tax payments and payments due
to creditors). Including this Deliveroo now forecasts it will have £[] of cash in June 2020 as opposed to the
£[] it previously forecast including deferred liabilities. The cash improvement between the two forecasts due to
deferred liabilities is £[].
120
start of the Coronavirus (COVID-19) crisis, with a significant decline in orders.
The outlook for online restaurant platform orders was very uncertain at that
time, and Deliveroo had (on the basis of the information that was then
available to it) forecast a prolonged period of lower revenue generation. Since
then, it has experienced a recovery in orders, which in April 2020 exceeded its
forecast at the start of the Coronavirus (COVID-19) crisis, and in May and
June 2020 exceeded its pre-Coronavirus (COVID-19) February Budget. This
resulted in a reduction in losses and the business being cash accretive over
the period from April to June 2020 which contributed to the cash balance at
the end of June 2020 substantially exceeding the forecast prepared at the
start of the Coronavirus (COVID-19) crisis in March 2020.
6.32 Deliveroo has stated that it does not expect this increase in profitability to be
sustained in the second half of the year. It expects the []. It also states that
the situation remains highly uncertain and that its financial position could be
negatively impacted by further lockdowns.
6.33 We note, however, that the expected net impact of these changes has already
been taken into account within Deliveroo’s own forecasts, which ultimately set
out:
(a) A considerable improvement in Deliveroo’s cash position (ie an increase
in its cash position, excluding deferred liabilities, in June 2020 from £[]
to £[]).
(b) A substantial reduction in losses (ie with actual losses for April, May and
June 2020 being £[] less than previously forecasted).
(c) A full year EBITDA-capex [] – which is better than the February 2020
forecast [].
(d) A positive forecast cash balance of £[] in December 2020, excluding
[],
255
exceeding the February 2020 budget forecasts.
Parties’ submissions
6.34 Since the publication of the April Provisional Findings, the CMA has sent a
number of requests for information to Deliveroo
256
and has had calls with
255
In assessing the counterfactual, the starting point is the situation pre-merger. Any transaction specific
elements such as [] must be removed from this analysis.
256
We sent requests for information (RFIs) to Deliveroo on 11 May 2020, 13 May 2020, 26 May 2020 and 17 July
2020.
121
Deliveroo and its advisors to discuss its financial situation.
257
In addition,
Deliveroo has made submissions to the CMA, including its response to the
Revised Provisional Findings.
258
6.35 Through these exchanges Deliveroo provided us with updated submissions
on its current financial position, as well as its actual and forecast trading
performance based on its updated experience of the pandemic. It has also
provided us with a reconciliation to previous forecasts and an explanation of
the key drivers behind the variances. This has been accompanied by updated
advice from Deliveroo’s external legal, financial and insolvency advisors.
These are examined in more detail below. In addition, Deliveroo has provided
us with an updated submission outlining its position in relation to the
counterfactual, which is also reflected in Deliveroo’s response to the Revised
Provisional Findings.
259
6.36 Deliveroo has submitted that any recovery in its financial performance is
irrelevant to the CMA’s assessment of the relevant counterfactual in this case.
This is because, in Deliveroo’s view, absent the Transaction, its funding need
would have arisen prior to the time that the recovery began, but after the
adverse impact on investor sentiment that occurred as a result of Coronavirus
(COVID-19).
260
6.37 In Deliveroo’s view, absent the Transaction:
(a) Deliveroo would have obtained cash funding of $[]–$[] million (as
opposed to $[] million under the Series G round);
(b) Deliveroo would have needed to raise additional funds in Q1 2020 when
the effects of Coronavirus (COVID-19) first appeared (and after the
adverse shift in investor sentiment);
(c) Deliveroo would have likely exited the market without additional funding;
and
(d) there would have been no likely sources of additional funding.
261
6.38 In its response to the Revised Provisional Findings, Deliveroo also submitted
that the CMA’s analysis of the amount of alternative funding Deliveroo would
257
We had a call with Deliveroo to discuss its financial position on 18 May 2020, and had a call with Deliveroo’s
advisors, [], on 9 July 2020 to discuss Deliveroo’s financial situation.
258
Deliveroo response to the Revised Provisional Findings, 10 July 2020.
259
Deliveroo response to the Revised Provisional Findings, 10 July 2020.
260
Deliveroo response to the Revised Provisional Findings, 10 July 2020, paragraph 5.13.
261
Deliveroo response to the Revised Provisional Findings, 10 July 2020, page 1.
122
have received at the time of the Series G funding round in the Revised
Provisional findings was inconsistent with its analysis in the April Provisional
Findings and that the CMA had placed insufficient weight on updated advice
from Deliveroo’s [] that it would still be facing [] in the absence of positive
provisional findings.
262
In addition, it submitted that the examples given by the
CMA as evidence of an improved funding environment concerned companies
either operating in different jurisdictions from Deliveroo or in different funding
environments.
Third party submissions
6.39 We also received a number of submissions from third parties in response to
the April Provisional Findings and gathered further evidence on the market. In
particular:
(a) Whilst the initial impact of the Coronavirus (COVID-19) pandemic on
online restaurant platforms, particularly in the last two weeks of
March 2020, was severe, some third parties submitted there had
subsequently been a [] recovery.
263
This was due in part to an increase
in the number of independent restaurants on these platforms, which offset
the impact of the closure of many chain restaurants. In support of this
position, one third party provided evidence showing a recovery in its order
volumes after a [] decline during the last two weeks of March 2020.
(b) Another third party submitted (on 18 May 2020) that, whilst it was still
experiencing relatively depressed order volumes, [] May 2020, []
June and July 2020.
(c) One third party submitted that other countries that had eased lockdown
measures before the UK had observed an improvement in online
restaurant delivery platform orders as restaurants re-opened.
264
(d) Third parties submitted that demand for takeaway delivery food has
increased due to the Coronavirus (COVID-19) crisis as consumers moved
from dining out to dining at home.
265
Third parties also identified
increased demand for online grocery deliveries.
266
262
Deliveroo response to the Revised Provisional Findings, 10 July 2020.
263
Just Eat Takeaway.com response to the April Provisional Findings, 11 May 2020, paragraph 21.
264
Just Eat Takeaway.com response to the April Provisional Findings, 11 May 2020, paragraph 20.
265
Just Eat Takeaway.com response to the April Provisional Findings, 11 May 2020, paragraph 17; Company D
response to the April Provisional Findings, 11 May 2020, paragraph 10.
266
Just Eat Takeaway.com response to the April Provisional Findings, 11 May 2020, paragraph 28.
123
(e) Third parties also noted that listed companies operating in the takeaway
delivery sector had outperformed the market since March 2020,
267
and
that demand for online restaurant platform apps had increased
significantly.
268
6.40 In addition, third parties provided evidence of an improvement in the funding
markets available to companies such as Deliveroo and that there was a wider
range of investors available.
269
They referred, in particular, to successful fund-
raising rounds announced following the April Provisional Findings by Just Eat
Takeaway.com and Wolt Group.
270
CMA assessment of the Deliveroo counterfactual
6.41 In the April Provisional Findings we assessed whether Deliveroo should be
considered to be an exiting firm based on the evidence available at that time
of the impact of the Coronavirus (COVID-19) pandemic. In keeping with our
duty to assess any additional evidence that comes to light during the course
of our investigation, and the evolving impact of the pandemic on Deliveroo’s
business, we reconsidered that assessment in light of the updated evidence
we have received following publication of the April Provisional Findings.
Limb 1: Would Deliveroo have exited absent the Transaction?
6.42 Our assessment follows the same structure set out in the April Provisional
Findings:
271
(a) Our assessment of whether Deliveroo would have exited absent the
Transaction.
(b) Whether the effect of Coronavirus (COVID-19) means that Deliveroo is
now likely to exit (excluding the merger-specific effects of the
Transaction).
(c) Whether Deliveroo would still have exited if one of the other available
sources of funding had proceeded.
267
Company D response to the April Provisional Findings, 11 May 2020, paragraph 8.
268
Just Eat Takeaway.com response to the April Provisional Findings, 11 May 2020, paragraphs 24–25 and
graphs.
269
Company D response to the April Provisional Findings, 11 May 2020, paragraph 16.
270
Just Eat announced it had raised €700 million worth of equity (€400 million) and convertible bonds
(€300 million) on 23 April 2020. See Takeaway.com Press releases, Just Eat Takeaway.com successfully raises
EUR 700 million through an accelerated bookbuild offering of new shares and convertible bonds and Bloomberg,
15 May 2020, Wolt Draws Goldman Investment in $108 Million Funding Round.
271
April Provisional Findings, 16 April 2020.
124
Would Deliveroo have exited absent the Transaction?
6.43 In the April Provisional Findings, we provisionally concluded that Deliveroo
would not have been an exiting firm absent the Coronavirus (COVID-19) crisis
because alternative funding would have been available at the time the
Transaction was agreed and Deliveroo would have expected to continue to be
able to raise subsequent funding.
6.44 This remains our conclusion.
Effect of Coronavirus (COVID-19)
6.45 In the April Provisional Findings, on the basis of the evidence available to us
at that time we provisionally concluded that, whilst the situation was inherently
uncertain, it was clear that the Coronavirus (COVID-19) crisis was having a
significant negative impact on Deliveroo’s business and cashflows, and this
was forecast to continue. We now consider whether more recent evidence of
the impact of Coronavirus (COVID-19) has changed this position.
6.46 In light of the cashflow forecasts and financial information provided by
Deliveroo at that time, and the advice received by its directors, we considered
that, without an assurance that it would receive additional funds in the
immediate future, Deliveroo’s directors would have been obliged to declare it
insolvent (and that Deliveroo would have exited the market). This reflected
forecasts of a significant decline in Deliveroo’s financial position compared to
its original budget, such that it would have run out of cash in [].
6.47 At the time of the April Provisional Findings, the Coronavirus (COVID-19)
crisis was rapidly developing and the ongoing impact of the pandemic on the
online restaurant delivery platform market was uncertain. Third parties have
confirmed to the CMA that like Deliveroo they also experienced a negative
impact on their business in the UK at the start of the crisis. One third party
submitted it experienced a [] decline in the last three weeks of March 2020
followed by a recovery whilst another reported a continued decline through
April 2020 []. Since then we have received additional information on the
effects of the Coronavirus (COVID-19) crisis and the impact of this crisis has
become clearer. Deliveroo has similarly been able to revise its forecasts to
reflect the evolving impact of the pandemic.
6.48 We note, in particular, that Deliveroo’s updated forecasts indicate that it now
expects to have improved cashflows and profitability through 2020 and a
positive cash balance until at least [] (when it expects to have £[]
125
remaining cash
272
according to the latest forecast). The CMA considers that it
is likely that it would not now need to raise additional funding until at least []
assuming it takes no further cost reduction measures (for example by
restructuring its operations), which could extend this further.
6.49 We note the advice that Deliveroo received from its insolvency advisors on
21 May 2020, and the statements made by [] to the CMA during a
telephone call on 9 June 2020, regarding the need to declare insolvency if it
does not have a reasonable expectation of receiving funds in order to ensure
it minimises the loss to creditors.
6.50 Deliveroo submitted that we failed to take account of advice from its
insolvency advisors []. Deliveroo submitted that ‘[I]n the “factual”,
Deliveroo’s improved trading performance in Q2 2020 did not alter the fact
that absent positive [Provisional Findings] (…) it faced [] insolvency’.
273
Deliveroo submitted its insolvency advisors advised that it had faced []
insolvency not due to its cashflow position (which had improved significantly)
but due to a technical insolvency position in which its liabilities would exceed
its assets.
6.51 We do not consider that the insolvency advisors’ advice supports a conclusion
that in the counterfactual Deliveroo would exit the market. As Deliveroo’s
submission notes, this advice was in respect of the ‘factual’. In the
counterfactual there would be a number of significant differences. By far the
most significant liability that Deliveroo has is []. In the counterfactual where
Deliveroo received alternative equity funding this liability would not have been
present.
6.52 Furthermore, the CMA did take account of the evidence submitted by
Deliveroo, including the advice received from its insolvency advisors, []. In
assessing any statements we receive, the CMA looks at the relevant
evidence. In some cases, we did not consider that [] advice reflected all the
evidence we have seen, or the correct approach to assessing the
counterfactual for merger control purposes, and we took that into account in
determining how much weight to put on that advice.
6.53 In light of the updated forecasts we now have available, we consider that
Deliveroo appears to have sufficient cash to be able to meet the demands of
its creditors until [] (excluding the []). This is a similar position to the one
it predicted prior to the Coronavirus (COVID-19) crisis. As such, it appears
that the specific impact of this crisis on Deliveroo’s liquidity has been
272
This excludes the requirement to [].
273
Deliveroo response to the Revised Provisional Findings, 10 July 2020, paragraph 5.14.
126
significantly less severe than was forecasted at the time of the April
Provisional Findings. This is due to combination of factors.
(a) Part of the cash increase is driven by a partial recovery in orders and
increase in commission per order which were not anticipated (and, given
the uncertainty, could not, in our view, reasonably have been predicted) at
the time of the April Provisional Findings.
(b) In addition to the significant cost savings undertaken by Deliveroo at the
time of the April Provisional Findings, the longer cash runway now
available has allowed it to implement further cost savings such as
redundancies that would not previously have been feasible.
(c) Furthermore, Deliveroo has been able to defer further payments to
creditors and tax authorities.
6.54 Accordingly, on the basis of the evidence now available, we consider that the
Coronavirus (COVID-19) crisis is having a less significant negative impact on
Deliveroo’s business and cashflows than we concluded was the case at the
time of the April Provisional Findings.
Would Deliveroo have exited if one of the other available sources of funding had
proceeded?
6.55 In the April Provisional Findings, we concluded that, absent the Transaction,
Deliveroo would have received funding in around April 2019 and would be
likely to have continued its pre-merger plans. The result would have been that
at the start of 2020 when the effects of Coronavirus (COVID-19) first
appeared, Deliveroo would already have been close to a position where it
would need to raise additional funds, reflecting its early and loss-making stage
of development. We also concluded, based on the evidence we had available
at the time, that the sudden and severe financial decline forecast by Deliveroo
would require it to seek urgent funding and that it would be unlikely to obtain it
at the time.
6.56 However, as explained in detail above, the updated financial information we
have received from Deliveroo indicates that, although the Coronavirus
(COVID-19) crisis had an initial severe impact on Deliveroo, the overall impact
has not been as severe as anticipated at the time of the April Provisional
Findings (and is now not forecast to be as severe in future). This recovery in
Deliveroo’s performance, combined with an increase in deferred liabilities and
cashflow improvements, has significantly improved Deliveroo’s forecast
cashflows to June. As set out in detail above, Deliveroo now has an additional
£[] in cash excluding deferred liabilities and an additional £[] when such
127
liabilities are included. This improvement would provide it with more time to
seek additional funding in the counterfactual. Furthermore, this would also
allow it more time to rationalise its business operations, potentially extending
this cash runway further.
6.57 The Parties put to us that, in the most likely counterfactual, Deliveroo would
have succeeded in raising funds of $[]–$[] million and amended their
budget, with reduced spending, resulting in an estimated closing free cash
balance of c.£[] at the end of March 2020. They assert that this would mean
that they would have had to raise funds in Q1 2020. On this basis, they
submitted that any improvement in Deliveroo’s performance during the
Coronavirus (COVID-19) pandemic would only have occurred after the point
at which its insolvency position became critical and therefore that Deliveroo
would have likely exited the market (given that the further funding would not
have been available at that time).
6.58 We do not agree with Deliveroo’s submission that the current forecasts, which
indicate that the impact of Coronavirus (COVID-19) on Deliveroo’s cash
position will be less significant than had been originally forecast at the time of
the April Provisional Findings, are irrelevant to our consideration of the
counterfactual.
6.59 We note that our provisional view in the April Provisional Findings that
Deliveroo was likely to exit the market relied on predictions of the sudden,
serious and sustained impact of Coronavirus (COVID-19) on Deliveroo’s cash
flow, which we found would also have occurred in the counterfactual. In the
absence of evidence of such a severe sudden decline to Deliveroo’s cashflow
in the period from March to June 2020, we have no basis for finding that such
a decline in cash would have occurred in the counterfactual.
6.60 The current evidence shows that Deliveroo has been able to manage the
specific impact of Coronavirus (COVID-19), through a combination of
deferring liabilities, saving costs and improving profitability, in order to leave it
in a similar cash position to the one it forecast in its December 2019 and
February 2020 budgets. This improved position and longer cash runway
means that:
(a) Deliveroo now has a less urgent funding need than it would have based
on the forecast position at the time of the April Provisional findings.
(b) This would also provide it with more scope to restructure its operations in
order to increase its cash runway further.
6.61 We have no basis to consider that Deliveroo’s ability to manage the impact of
Coronavirus (COVID-19) would have been different in the counterfactual.
128
6.62 Deliveroo’s submission in respect of Limb 1 now appears to rest on the
submission that its ‘ordinary course’ funding need would have arisen in
March 2020 in the counterfactual (and alternative funding would not have
been available). We consider Deliveroo’s submissions on the amount of
alternative funding it would have raised and the amendments to its budget as
a result of this too speculative to form the basis for a conclusion that the most
likely counterfactual involves Deliveroo’s exit:
(a) Firstly, the evidence does not reveal the precise amount that would have
been raised by Deliveroo in the counterfactual, although we consider that
the available evidence shows that alternative funding offers were of a
broadly similar scale to the $[] million Series G funding amount. We
therefore consider that there is no reasonable basis to conclude that
Deliveroo would have been in a materially worse financial position in the
counterfactual than it is currently.
(b) We also cannot speculate on how precisely Deliveroo would have then
operated the business following this fundraising. Deliveroo’s submission
with respect to the continued relevance of the exiting firm counterfactual
would require us to assume that Deliveroo would have raised a lower
level of funds in early-2019, that having received that lower level of funds
it would have operated its business in entirely the same manner as it did
having received a larger level of funds, and that it would have taken no
steps prior to early-2020 to secure additional funding, resulting in the
business running out of funds and exiting the market in March 2020 at the
height of the Coronavirus (COVID-19) crisis. While that is clearly one
possible alternative outcome, we consider it is not a likely outcome.
6.63 Contrary to Deliveroo’s submissions,
274
this approach is consistent with the
approach taken in the April Provisional Findings. As set out in the April
Provisional Findings, Deliveroo ceased negotiations on alternative funding as
it became clear the Amazon deal would progress and we are unable to say
precisely how much alternative funding Deliveroo would have raised at that
time. This remains our view and, as set out in the April Provisional Findings,
as Deliveroo did not progress [] the options described above far enough to
understand the precise quantity or terms of the investment we are unable to
state with precision how much Deliveroo would have succeeded in raising.
275
6.64 Accordingly, we do not have sufficient evidence to conclude that the nature
and timing of Deliveroo’s funding needs in the counterfactual would mean that
274
Deliveroo response to the Revised Provisional Findings, 10 July 2020, paragraph 5.8.
275
April Provisional Findings, 16 April 2020, paragraph 4.31.
129
the most likely counterfactual involves its exit from the market in the way
suggested by the Parties.
Conclusion on the application of ‘Limb 1’ of the exiting firm analysis
6.65 As described above, our April Provisional Findings in respect of ‘Limb 1’ of the
exiting firm analysis were based on a finding that Deliveroo would have
experienced a sudden and severe financial decline as a result of Coronavirus
(COVID-19) that would have required it to seek funding urgently and would
have meant that it would not have sufficient cash runway to undertake and
benefit from restructuring. The evidence that we now have does not support
that finding. Although order volumes declined very significantly in March 2020
leading to a sharp deterioration in Deliveroo’s financial position, Deliveroo’s
cash position is significantly improved from that forecast in the documents
underpinning the April Provisional Findings, and in fact from that forecast in
Deliveroo’s pre-Coronavirus (COVID-19) budget, principally as a result of an
increase in revenues and gross margins (which was not foreseen at the time
of the April Provisional Findings), in addition to creditor deferrals described
above. We would expect the situation to be similar in the counterfactual.
6.66 Whilst we acknowledge that, in the counterfactual, Deliveroo would have
continued to operate in a way that was reliant on external fundraising, the
available evidence does not support Deliveroo’s submission that its exit would
have been imminent, absent the Transaction, in March 2020.
6.67 In light of the evidence that Coronavirus (COVID-19) has had a more limited
impact than expected on Deliveroo’s business, our view of the most likely
counterfactual is therefore that Deliveroo would not exit the market as a result
of the impact of the Coronavirus (COVID-19) crisis and would continue to
compete in the market, as it expected prior to the crisis and as described in
the Merger Notice.
Limb 2: Is there a substantially less anti-competitive investor or purchaser for
Deliveroo’s assets?
6.68 As our view is now that Deliveroo would not meet ‘Limb 1’ of the exiting firm
scenario, Deliveroo cannot be considered as an exiting firm for the purposes
of the counterfactual and therefore there is no need to consider ‘Limb 2’ and
‘Limb 3’ of the applicable test.
6.69 However, for completeness (and in light of the fact that Deliveroo remains
reliant on external funding), we also note that regardless of the alternative
investment in Deliveroo in the counterfactual, we would expect Deliveroo to
have made prudent plans to raise finance in good time (which is consistent
130
with its strategy to date). As noted above, at the time of the April Provisional
Findings and on the basis of the evidence we had at that time, we believed
that Deliveroo was facing a financial ‘cliff-edge’ as a result of the significant
impact of Coronavirus (COVID-19) with no evidence of an improvement in the
near future, which would have required funds to be raised in a very short
period of time, during the period that investor sentiment was weakest.
6.70 In light of the evidence we now have, we consider that Deliveroo’s cashflow
position (in both the factual and the counterfactual) has improved significantly
albeit with a short-lived deterioration at the start of the crisis from which it has
since recovered. Therefore, there is no basis, on the evidence available to us,
to suggest that its funding needs in the counterfactual would have been
materially different in nature and urgency from that envisaged prior to the
crisis.
6.71 Whilst we have received evidence from Deliveroo’s professional advisors that
a company in Deliveroo’s position would find it ‘highly challenging’ to raise
funding in March 2020 our view is that in the relevant counterfactual Deliveroo
would likely have taken action to ensure that it was both more attractive to
potential investors and to seek to avoid a ‘cliff edge’ in which it must secure
funding very urgently. Deliveroo could for example have rationalised some of
its operations to provide a longer cash runway and reduced its losses making
it more attractive to some investors. In addition, it would no longer have had
the urgent funding need previously envisaged in the April Provisional findings
due to Coronavirus (COVID-19).
6.72 There are also a number of factors that suggest that the funding environment
for Deliveroo has improved since the April Provisional Findings, all of which
would be equally applicable in the counterfactual. These include:
(a) the recent successful fund raising by Just Eat Takeaway.com,
276
Wolt
Group and DoorDash;
(b) the improvement in Deliveroo’s cash position compared to that originally
forecast at the start of the crisis; and
(c) the improved investor sentiment to this sector, as shown by the recovery
in share price of listed companies in this sector.
277
276
Just Eat Takeaway.com confirmed this was not contingent on phase 1 clearance.
277
Listed companies in this sector in the UK and abroad including Grubhub Inc (which Just Eat Takeaway agreed
to buy on 10 June 2020 (see Just Eat Takeaway.com to combine with Grubhub to create a leading global online
food delivery player)), Delivery Hero SE NPV, Just Eat Takeaway.com N.V and Domino’s Pizza Inc. The recovery
in Just Eat Takeaway.com N.V’s share price occurred prior to the announcement by the CMA of its phase 1
merger clearance and fundraising announcement on the 23 April 2020.
131
6.73 On this basis, while we have not been required to consider whether Deliveroo
would have been able to raise additional funds specifically to survive the
Coronavirus (COVID-19) crisis absent the Transaction, we note that the
available evidence indicates that there is now a materially greater likelihood
that alternative sources of funding would have been available to Deliveroo.
In addition, Deliveroo no longer faces (and would not in the counterfactual
have faced) such an urgent deterioration in its cash position as a result of
Coronavirus (COVID-19), providing more scope to manage its funding needs
through rationalising its operations.
6.74 As such, we consider that the evidence that the impact of Coronavirus
(COVID-19) on both Deliveroo and the funding markets has been more muted
than expected at the date of the April Provisional Findings means that the
most likely counterfactual is that Deliveroo would have continued to compete
in the market, and to raise funds to do so, in the same way as it anticipated
prior to the crisis.
Amazon counterfactual
6.75 As mentioned in paragraph 6.2, the CMA may also look at alternative
counterfactuals where one (or both) of the merging parties would have
entered or materially expanded its presence in a market absent the
transaction.
6.76 During its phase 1 investigation, the CMA assessed the impact of the
Transaction against an alternative counterfactual where Amazon re-entered
the supply of online restaurant platforms in the UK. In this context,
278
for our
phase 2 investigation, we examined whether the most likely scenario, absent
the Transaction, is that Amazon would choose to enter the supply of online
restaurant platforms in the UK.
6.77 Amazon is a large and complex international business, and we assessed
evidence from a variety of parts of its business, including documents prepared
by or presented to its senior executive team, the corporate development team,
the Amazon Prime and F3 teams (‘fresh food fast’, which includes Amazon
Fresh and Prime Now), which used to also encompass the Amazon
Restaurants business. We considered evidence from teams both in the UK
and overseas, particularly in the US, where most senior staff are located and
where Amazon has the largest and most developed market presence. We
278
The counterfactual used by the CMA may differ between phase 1 and phase 2 investigations because of the
different legal standards used. At phase 1, the CMA will assess the merger against the most competitive
counterfactual that it considers is realistic. At phase 2, the CMA must assess the merger against the most likely
counterfactual. See MAGs, paragraphs 4.3.5 and 4.3.6.
132
observed teams producing global strategy documents for the executive team
based in the US, which featured global strategies and aims for Amazon, albeit
incorporating some national specificities.
6.78 During our investigation we gained a detailed understanding of Amazon’s
overall strategy as a business and how it operates in practice. This has
provided important context to our assessment.
6.79 First, we learned that Amazon does not, in general, produce large volumes of
internal documents to inform its commercial strategy. Amazon told us it
encourages the [] rather than having ‘[]’, and that decisions are often
taken in meetings. Although we have observed documents produced for
discussion at meetings, Amazon told us that it does not typically make written
records that summarise the discussions, or circulate notes describing the
decisions and actions arising from these meetings. In some cases, Amazon
has told us that it had not retained the documents we requested to inform our
assessment, including the emails sent or received by [], who led Amazon
Restaurants and currently leads F3.
279
These factors made it difficult to fully
understand Amazon’s rationale for its decisions in some instances and the
lack of clear, conclusive evidence on certain points has meant we have relied
on a range of documents, considered in the round, to inform our view. It has
also meant that we have not been able to conclude that a lack of detailed
documentary evidence on particular issues (such as Amazon’s future strategy
and plans with regard to restaurant delivery) means that such issues and
related decisions are not being considered by Amazon, and that such issues
are not important to it.
6.80 Second, Amazon has told us that it takes a ‘test and learn’ approach,
innovating or expanding its offering regularly.
280
Amazon told us that it [].
We noted documents referring to ‘two-way door’ decisions;
281
that is, the
ability to reverse decisions. Of one decision to withdraw a service, Amazon
told us ‘[]’. We therefore consider that previous business decisions taken by
Amazon should not be considered to irrevocably fix its intended course in a
given market. This type of approach by Amazon has informed how we assess
the decisions Amazon makes, including the closure of Amazon Restaurants.
6.81 In order to assess the likelihood of Amazon re-entering the supply of online
restaurant platforms in the UK, we have assessed Amazon’s incentives and
intention to re-enter, primarily by examining evidence from Amazon’s internal
279
Amazon told us [].
280
For example, at the main party hearing with Amazon on 31 March 2020 [] told us that ‘[]’.
281
For example, in an Amazon internal email, provided to the CMA on 2 April 2020, [] questions whether
closing Amazon Restaurants in the US is a ‘one-door decision’ and considers Amazon would ‘probably’ be able
to re-enter in the future.
133
documents. We then assessed Amazon’s ability to re-enter, again by
examining Amazon’s internal documents and looking at evidence from third
parties. As noted above, the counterfactual is an analytical tool and we are
not required to separately consider, on the balance of probabilities, whether
Amazon has each of (i) the intention, (ii) the incentive and (iii) the ability to re-
enter. In our assessment, we have considered all of the evidence in the round
(including Amazon’s commercial strategy, the international nature of its
business and the nature of the online restaurant platform market) to determine
the most likely counterfactual.
6.82 Most of our evidence gathering was completed prior to the outbreak of
Coronavirus (COVID-19) and, in particular, our assessment of Amazon’s
incentives and intention to enter relies on evidence from internal documents
produced prior to Coronavirus (COVID-19). Where possible we have sought
to consider what impact Coronavirus (COVID-19) might have had on both the
overall market for restaurant platforms in the UK, as well as on the likelihood
of Amazon’s re-entry into this market.
Parties’ submissions
282
6.83 The Parties submitted that Amazon is not a likely re-entrant into online
restaurant food delivery in the UK, pointing to its failed entry through Amazon
Restaurants (which operated in the UK between 2016-2018) as evidence that,
even with Amazon’s resources and recognition, it is not easy to execute such
a business.
6.84 They submitted that restaurant delivery is not [] to Amazon’s UK strategy
(otherwise it would not have shut down its offering in 2018), that it is not
looking to [], and its general interest in online food delivery does not
support a conclusion that Amazon would have re-entered that space in the
UK.
283
They submitted that restaurant delivery would be ‘[]’ for Amazon and
that Amazon’s modelling in the context of [] showed a [] Net Present
Value (NPV) to Amazon of building a restaurant delivery platform. They
submitted that Amazon does not have a compelling commercial imperative to
re-enter restaurant delivery in the UK, as shown by the fact that it decided
[].
6.85 The Parties submitted that there are material barriers to entering restaurant
food delivery in the UK, including significant technological barriers and
investment needed in attracting restaurants, couriers and consumers and
282
Parties’ Initial Submission, 24 December 2019.
283
Amazon response to the Revised Provisional Findings, 10 July 2020, page 2.
134
developing an appropriate logistics model, and that Amazon is ‘not particularly
well placed’ to overcome these.
6.86 Finally, the Parties submitted that a 16% investment in Deliveroo would not
materially impact Amazon’s incentives on whether (or how) to re-enter
restaurant delivery in the UK. For the purpose of assessing what the most
likely counterfactual is, we need to consider what we believe the most likely
scenario is absent the Transaction, and therefore the fact that the Transaction
involves the acquisition of a minority shareholding is not relevant to this
assessment. The impact of the Transaction (in particular it being a minority
shareholding as opposed to a full merger) on Amazon’s incentives to re-enter
is assessed and discussed in the competitive assessment, from
paragraph 7.1.
CMA assessment: Amazon’s incentives and intention to re-enter
6.87 In order to assess Amazon’s incentives and intention to re-enter the supply of
restaurant platforms in the UK, we reviewed a number of Amazon’s internal
documents and emails from various business areas in order to gain as full an
understanding as possible of Amazon (as detailed in paragraph 6.78), and its
incentives and intentions in this market. We requested detailed explanations
of Amazon’s decision-making process and the related documents, including
details of the timing of discussions and decisions, the key senior staff involved
in these discussions and decisions, and other contextual information from
Amazon. This built on the evidence gathered and analysis carried out by the
CMA during its phase 1 investigation, which included the use of the CMA’s
formal powers to interview senior Amazon staff. We gathered evidence from
numerous third parties including restaurants, competitors and professional
analysts.
6.88 In this section, we first discuss evidence relating to Amazon’s current
strategies for Prime and its food offering within this, then consider evidence
relating to its previous attempt at supplying a restaurant platform, Amazon
Restaurants, and then finally evidence relating to its interest and subsequent
investment in Deliveroo.
Amazon Prime
6.89 Promoting and growing Prime is very important to Amazon, and it has
successfully continued to grow Prime membership globally in recent years.
Amazon’s CEO has described this as follows: ‘If you look at Prime Members,
they buy more on Amazon than non-Prime members…once they’ve paid their
annual fee, they’re looking around to see how can I get more value out of the
135
program? So they look across more categories, they shop more…a lot of their
behaviors change…it really is a flywheel’.
284
6.90 After the US, the UK is Amazon’s [] largest territory in terms of Prime
subscription revenue ([]) and Prime household penetration ([]). The UK
has the [] largest total number of Prime customers ([]).
285
This number
has grown rapidly in recent years, from [] in 2016 to [] in 2018. It is
forecast to continue growing and [].
6.91 An Amazon operational plan (OP) for Prime states that ‘[]’, and that
Amazon []. []. Amazon has invested heavily in its Prime proposition in the
UK, in particular by recently adding Premier League football and other live
sports to Prime Video, where it expects to ‘[]’ for the UK.
6.92 Amazon’s internal documents provide evidence of customer loyalty to Prime.
In 2014, Amazon increased the price of UK Prime by []. Associated with this
price change was an [].
286
[]. The fact that Amazon has not increased the
price of Prime in the UK in recent years, [], indicates that Amazon is not
solely focussed on short-run profitability at this stage. The available evidence
indicates that growing its Prime membership is the more important objective
for Amazon at this time. [], indicating this is a particularly important aspect
of Amazon’s commercial strategy.
6.93 A third party told the CMA that ‘Prime is a bundle of unrelated products,
combined to protect Amazon’s most valuable customers’ and that ‘focusing on
a single specific segment misses how Amazon exercises its market power’. It
said that ‘Prime … was designed to create a moat around Amazon’s best
customers. People feel they have committed to Prime, and want to get value
out of this – it’s a behavioural issue’. Other third-party submissions also
commented on the power of Prime for Amazon, expressing concerns about
the way Amazon bundles different services into Prime, making it more difficult
to compete against.
287
284
Jeff Bezos in a 2016 Recode interview, as quoted on Forbes.com. The term ‘flywheel’ is used at Amazon to
describe something similar to a virtuous cycle, which powers the business. A service such as restaurant delivery
could attract more users to Amazon, who then use the other services Amazon offers, such as Amazon Video or
the Amazon marketplace. Each benefit is seen as contributing to the flywheel. Amazon also uses this in
reference to costs. For example, lower prices lead to more customer visits, more customer visit increases the
volume of sales, and that results in more commission-paying third-party sellers to the site. [].
285
Figures in this paragraph are all taken from [].
286
The equivalent figure for new customers was [].
287
The CMA considered a conglomerate effects theory of harm during its phase 1 investigation but did not find a
realistic prospect of an SLC as a result of the Transaction. This is discussed again in Chapter 9, given it was an
area we received a number of submissions on.
136
Offering restaurant delivery as part of Prime
6.94 Amazon told the CMA that it has guidelines to identify and evaluate new
Prime benefits. These guidelines state that ‘[]’ and that it []. []. Amazon
submitted that any new benefit would need to [].
6.95 Food in general would appear to be an area which Amazon sees as high-use
and of value to its Prime customers, and capable of generating a large
incremental spend from customers – the most recent OP document for Prime
in the UK notes ‘[]’.
6.96 Restaurant delivery is seen as a useful benefit for Prime. First, this was
shown by Amazon’s expectations of Amazon Restaurants. As described in
paragraph 6.133, it was seen as a [], and was expected to []. The 2019
Amazon Restaurants OP1 document stated that Amazon Restaurants was
‘[]’. [].
288
6.97 Amazon submitted that the fact that these aspirations were not achieved and
that Amazon Restaurants was closed by senior leadership shows []. It also
submitted that Amazon Restaurants was so [] that it risked []. We
consider that this factor may explain why Amazon decided to close Amazon
Restaurants but the closure does not mean that the expected benefits no
longer exist and that Amazon would not want to offer restaurant delivery if it
could do so successfully. It follows that a high-quality restaurant delivery
service could enhance relationships with Prime customers. In addition, as
explained to us by Amazon and commented on by third parties, Amazon
regularly tests propositions, learns from these and innovates in this manner.
We consider the withdrawal of the service does not mean that the issue is
closed and that it would not re-enter in the future.
6.98 Second, the attraction and benefit of restaurant delivery [] for discussion
with Jeff Bezos and other members of the senior leadership team. This
document was used to discuss whether [] Deliveroo at a meeting on
12 March 2019, where the decision to [] consider investing in Deliveroo was
taken. The document makes the following observations:
(a) With respect to the ‘flywheel accelerants’, the document stated Amazon
could expect []. The document noted ‘[] (DSI-CP
289
impact of [])
[]. []’.
288
Amazon’s response [] states ‘[]’.
289
Downstream impact on contribution profit.
137
(b) Amazon forecast [].
(c) In this document, Amazon uses its experience from Amazon Restaurants
in the US to infer the [].
290,291
(d) It states that the total downstream impact on contribution profit in the UK
from this effect would be £[]. In revenue terms, across the same
customer base, this works out at £[].
292
6.99 Amazon submitted that it chose not to []. It also submitted that the DSI
analysis provides evidence of [], not []. Finally it submitted that this
analysis was relegated to an appendix [], displaying a lack of reliance on
the figures. We consider that the analysis [] is helpful evidence in
understanding how Amazon views the market and how restaurant delivery
would fit in its wider business. Second, we note that the document makes no
mention of the DSI impact being ‘[] not []’ or caveats to this effect. Finally,
we note that the findings of the DSI analysis are included in the main
summary of the document as the second reason for why Amazon []. The
detail is included in an appendix, as appears to be the standard format of
Amazon internal documents.
6.100 Even following the decision in March 2019 to look to [], Amazon considered
the idea of [], indicating that it sees restaurant delivery as an attractive
benefit to its customers. Amazon told the CMA that this was a ‘back of the
envelope idea’ and was not reflective of leadership views, and that the sender
himself was ‘a little sceptical of value add and likely impact’. But [] (VP,
Corporate Development) replied that []. In another email referring to
promoting Deliveroo Plus to Prime members, [] (VP, F3) []. Evidence
therefore indicates that these ideas were supported by senior leadership and
discussed within the business.
6.101 The Parties submitted that only [] placed an order with Amazon
Restaurants. Figures provided by Amazon showed that during the time
Amazon Restaurants operated in the UK it received orders []. As we explain
in more detail later in this chapter (at paragraph 6.141 onwards), our view is
that the evidence shows Amazon Restaurants was not well or fully executed
in the UK – with Amazon [] in the business, and conducting []. In
addition, we note that Amazon Restaurants was only available to customers in
selected Greater London postcodes, representing only a small proportion of
total Prime members. We consider these factors may explain why the take-up
290
Using an exchange rate of 0.786130328 for 4 December 2018:
https://www.xe.com/currencytables/?from=USD&date=2018-12-04.
291
This figure comes from Amazon’s restaurants experience in the US, where customers ‘[]’.
292
[]
138
of restaurant delivery was low for Prime customers in the past, but that it does
not necessarily follow that restaurant delivery is not a benefit that Prime
customers would value. This is illustrated by [] (which generally highlights
the demand from Amazon Prime customers for restaurant delivery) and is
consistent with the views of Amazon’s senior management.
293
6.102 As mentioned in paragraph 6.90, Prime is a very popular subscription product
in the UK, with Amazon forecasting that it will have [], higher than []. In
addition to the UK being attractive to Amazon because of the size of Prime, it
is also an attractive country for online restaurant platforms. It is one of the
largest markets for restaurant delivery in the world and is the largest in
Europe. Past trends show UK household expenditure on takeaway food
continues to grow, and people are continuing to move to online ordering. Third
parties told us that the UK was an attractive market because of its size and
frequency of orders, high proportion of urban population and high GDP, and
because of existing merchant and courier demand. These factors would
increase Amazon’s incentive to operate an online restaurant platform in the
UK. One overseas restaurant delivery business who told us the UK was an
attractive market provided evidence to us showing it had recently considered
entering the UK (although had decided against it at the current time).
6.103 On the other hand, overseas competitors told us that it would be hard to enter
the UK because of the existing players, and so less attractive. In its response
to our Revised Provisional Findings, Deliveroo submitted that the fact that
overseas competitors saw the UK as competitive and would be disinclined to
enter was inconsistent with our finding that Amazon would be likely to re-
enter.
294
We consider that there are differences between the barriers to entry
faced by Amazon and overseas competitors, particularly in terms of customer
acquisition (described below), which may lead to different incentives. Some
overseas operators appear to prefer to focus on emerging markets where they
may have a first mover advantage and customer acquisition may be less
expensive, for example Glovo has been expanding in Latin America and to
countries such as Georgia and Kazakhstan. In this regard, we consider
Amazon would have a significant advantage over other operators in the UK as
it can benefit from having existing relationships with millions of customers in
the UK, including engaged customers through Prime, who are already part of
Amazon’s flywheel. Amazon would be able to advertise to customers at a
lower cost than rivals, particularly in the UK given the popularity of Prime in
the UK. We consider this makes the UK more attractive to Amazon as a
293
A survey in one of Deliveroo’s internal documents show that []% of Deliveroo Plus customers subscribe to
Amazon Prime.
294
Deliveroo response to the Revised Provisional Findings, 10 July 2020.
139
country to enter restaurant delivery than it is for overseas restaurant delivery
competitors. On the basis of the evidence available about its broader
business strategy (eg its focus on growing Prime as opposed to short-run
profitability) and the patience of Amazon as both an operator and an
investor,
295
we also consider that Amazon may have a different time horizon
for profitability compared to other potential entrants and the financial
resources to support this.
6.104 Coronavirus (COVID-19) has had a large impact on restaurants, causing
many to temporarily close or move to providing food for takeaway or delivery
only. These shifts, as well as a potential reduction in household expenditure
on restaurant food (because of reduced disposable income) could lead to
significant permanent changes in the market, including restaurant closures.
Deliveroo submitted that it expects up to half of its independent restaurant
partners to become insolvent when the government support packages stop,
and that independent restaurants will be disproportionately affected compared
to chain restaurants. On the other hand, restaurants are beginning to re-open
and adjust their business models, with delivery forming a more important part
of their business. Food delivery has become increasingly important for
consumers. Data from Deliveroo indicate a recovery in order levels during
April 2020, with order levels exceeding those during April 2019 (see further
discussion in Chapter 2). The long-term impact of Coronavirus (COVID-19) on
the restaurant delivery market in the UK is uncertain. It is not clear, based on
the evidence available at this time, whether Coronavirus (COVID-19) will
make the UK a more or less attractive country in which to operate as an
online restaurant platform. As we therefore cannot determine with sufficient
certainty, for the purposes of our counterfactual assessment, whether
Coronavirus (COVID-19) is likely to have an enduring negative impact on the
attractiveness of the UK market, we consider that it is reasonable to rely on
the broader body of evidence to date showing the potential attractiveness of
the UK market to Amazon (notwithstanding recent market fluctuations).
Amazon’s launch of its own restaurant food delivery service: Amazon India
6.105 Amazon told the CMA in August 2019 that it had no internal documents that
assessed the possibility of acquiring a restaurant delivery service in the UK,
US or elsewhere, []. Amazon did submit that given the broad and global
nature of Amazon’s business, ‘[]’. In response to further statutory
information requests, it later confirmed in February 2020 that it was launching
Prime Food India, a restaurant delivery service for Prime members. According
295
April Provisional Findings, 16 April 2020, paragraph 2.54(b).
140
to the dates of Amazon India documents later provided to us, Amazon had
been considering launching restaurant food delivery in India since [].
6.106 India is a rapidly growing country for Amazon in terms of Prime membership,
which Amazon forecasts to []. Amazon appears to have invested
significantly ($6.5 billion) in expanding in India in recent years.
296
6.107 With respect to offering restaurant delivery in India, Amazon submitted that
there are significant differences between the regulatory and commercial
conditions in India compared to the UK and its assessment of industry
conditions, competition and business opportunities in India is ‘completely
different’ to that of the UK.
6.108 We acknowledge that there are clear differences between the business
environment in the UK and India. We consider, however, that the evidence
shows that Amazon has a global strategy (which applies in both the UK and
India) around growing and promoting Prime membership with the intention of
expanding its position in groceries and in fast delivery across various
countries. In addition, we note the factors considered by Amazon in launching
a restaurant delivery service in India were similar to those considered by the
Amazon Restaurants team in assessing why offering restaurant delivery was
beneficial for Amazon in the UK and the US (see discussion from
paragraph 6.125). Amazon is using technology previously developed by its
Amazon Restaurant offering in the US/UK in its India offering, showing the
potential utility of technology and expertise gained in India for its global
ambitions in restaurant delivery, including the UK.
297
Therefore, we consider
the strategic thinking and approach taken in India is relevant evidence in
assessing Amazon’s broader global strategy in relation to online restaurant
platforms, including in the UK.
6.109 An Amazon India paper discussing the launch of restaurant delivery from
November 2018 features notes from a key discussion on the project, including
comments from [] (SVP, Amazon India), which state that there are three
key objectives for launching restaurant delivery: ‘[]’. The document goes on
to discuss the way Amazon India should market the offering, including that it
should [].
296
BBC, 21 May 2020, Amazon trials online food delivery in India.
297
Amazon submitted that it was not credible that the Transaction could preclude it from ‘experimentation at a
global level and even less credible that Amazon would decline to apply learnings and technologies from other
countries of operation to the UK’.
141
6.110 In a subsequent Amazon India paper, dated 31 January 2020, Amazon India
discusses []. This document mentions that it ‘[]’. The document also has
a section on how it plans to [].
6.111 Finally, we note a pattern of restaurant delivery players expanding into
offering fast delivery of other products in various markets across the world,
not just in India. For example, DoorDash offers fast grocery delivery in the US,
and Delivery Hero has publicly stated it intends to challenge Amazon with fast
delivery of consumer items such as groceries.
298
Conclusion on Amazon Prime
6.112 Offering fast delivery of a range of products, including restaurant food, is also
seen as a way to enhance the value of Prime, which the evidence shows is
important to Amazon’s overall global corporate strategy. Amazon has
considered restaurant delivery on a global level in the context of its desire to
attract and retain the customers of its Prime subscription service and to be
known for fast delivery. Amazon has a global strategy, as explained below, of
expanding its grocery offering and increasing the speed of delivery.
Restaurant delivery through Amazon Restaurants was expected to play an
important role in this. This is also supported by the evidence in relation to
Amazon India, where restaurant delivery is seen as an important benefit for
customers, as well as a way to develop operational capacity in fast delivery
and also as a [].
Amazon food strategy
6.113 F3 is an Amazon business area which used to also include Amazon
Restaurants. F3 operates internationally and is focused on developing
Amazon’s online grocery offering including through the roll-out of its ultra-fast
grocery (UFG) plan. This strategy is intended to []. During an interview with
the CMA, [] (VP, Corporate Development) stated ‘[]’. Food is one of the
largest areas of expenditure for UK consumers, and an area in which Amazon
has so far not managed to penetrate, despite its wide-ranging activities in the
delivery of other items, such as consumer electricals and baby products.
299
6.114 Amazon submitted that it closed Amazon Restaurants [], and pointed to the
F3 OP document which details this and was produced a few months after the
decision to close Amazon Restaurants was made.
298
Bloomberg article, 30 July 2019.
299
See Sainsbury’s/Asda, Final Report, 25 April 2019, paragraph 11.38, showing Amazon has a [0-5]% market
share in online delivered groceries.
142
6.115 The F3 strategy documents show a clear focus on grocery delivery, including
how Amazon intends to offer wide selection and ultra-fast delivery, stating
Amazon will offer ‘[]’. The documents do not generally discuss restaurant
delivery. This document does mention that the UK offering is [], which in the
US includes restaurant delivery. The appendix to this document states ‘[]’.
We note that when this document was produced ([]), Amazon Restaurants
was still operating in the US. Amazon submits it is therefore logical that the
US vision would include restaurant delivery but that this only applies to the
US. We consider that the inclusion of restaurant delivery in its grocery
strategy shows that Amazon continued to see a role and benefit of restaurant
delivery in its global food strategy despite having closed Amazon Restaurants
in the UK.
6.116 As stated above, Amazon submitted that expanding in online groceries being
F3’s priority is a reason it would not enter restaurant delivery. We do not
believe that developing a restaurant delivery service needs to be the key
priority of Amazon for us to believe that it could re-enter. Given the size and
resources available to Amazon, it could easily pursue multiple strategies and
projects concurrently, provided it had sufficient incentives to do so.
6.117 As part of its US grocery offering, Amazon has rolled out the use of [].
Amazon’s ambitions and the developments it plans in grocery delivery would
help it to overcome one of the main challenges it faced when operating
Amazon Restaurants. Extending this to include restaurant food would be a
natural expansion as Amazon would have one of the key building blocks of a
successful online restaurant platform. Third parties who offer on-demand
delivery have told us that offering delivery across different categories of
product (such as restaurant food and groceries) helps to provide density
throughout the day for the courier fleet and improves unit economics.
6.118 Several internal emails from within F3 refer to the importance of offering
restaurant delivery as part of Amazon’s food strategy, and in particular [],
including:
(a) An email chain where [] (VP F3) disagrees with the decision to close
Amazon Restaurants, where she lists the benefits including differentiating
Amazon’s food offering in terms of speed and breadth and being an
important Prime benefit to customers in a growing market.
(b) One email from February 2019 setting out the recommendation of []
(VP F3), as well as [] (who was General Manager of Amazon
Restaurants), who ‘[]’.
(c) Another email from [] discussing whether [].
143
6.119 These are supported by emails from the corporate development team that
show consideration of wider benefits to Amazon from [] Deliveroo, [].
One email, written by [] (VP Corporate Development) states that ‘[]’.
300
[] (Senior Manager, Corporate Development) agreed with [] (VP
Corporate Development) and [] (SVP Business Development) and stated
that they ‘[]’.
6.120 Amazon’s strategy in India provides evidence of how restaurant delivery fits
into Amazon’s overall global food strategy. Amazon is currently trialling
restaurant delivery in India while investing significantly in expanding its
presence. In a press article reporting on the trial, Amazon stated that
‘Customers have been telling us for some time that they would like to order
prepared meals on Amazon in addition to shopping for other essentials. This
is particularly relevant in present times as they stay home safe. We also
recognise that local businesses need all the help they can get’.
301
6.121 Amazon’s internal documents show the thinking behind its investment in
restaurant delivery in India, mirroring the F3 reasoning, including that
restaurant delivery offers differentiation to Amazon’s service and will allow
Amazon to develop logistical expertise that it can apply in its wider business.
(a) In one email exchange from June 2019, it is noted that Swiggy is [].
(b) A PRFAQ (Press Release and FAQ) document, produced by the Amazon
India team and discussed with [] (SVP, Amazon India), explains how
[], and how the development of logistics for this business can be used
more widely in Amazon in the future. It states ‘[]’.
6.122 Amazon submitted that the developments in India are completely separate
and specific to the environment in India and that []. We consider that
whether or not [] is only of limited relevance given that all Amazon business
units ultimately form part of a single firm under the direction and control of the
same senior management . Indeed, [], many of the expected benefits and
the reasoning for offering restaurant delivery in India, for example [], are
similar []. This is consistent with the position that these ambitions are
globally important goals ([]) and that if restaurant food delivery is an
important element in achieving these goals in India, there is no reason to think
that it would not be so in other geographies.
6.123 As described above, expanding a fast grocery delivery service is a clear area
of focus for the F3 team, both internationally and in the UK. We consider this
300
DoorDash and Postmates are US restaurant delivery services who also offer fast delivery of groceries. Uber
has recently agreed to acquire Postmates (on 6 July 2020) (see Uber to Acquire Postmates).
301
BBC, 21 May 2020, Amazon trials online food delivery in India.
144
aim is unlikely to be changed following the spread of Coronavirus (COVID-19)
and the increased demand and need for grocery delivery. Responding
effectively to the impact of the virus on F3 operationally is likely to occupy
significant senior executive time and effort. Amazon submitted that as a result
of the abrupt increase in demand for its grocery services as a result of
Coronavirus (COVID-19) it faced [] and was required to make a number of
operational and technical changes to maintain its customer experience. This
could mean the [] for launching new services is less than that prior to
Coronavirus (COVID-19). At the same time, the delivery of food, including
restaurant meals, has become increasingly important to customers at the
moment (as acknowledged by Amazon in the quote in paragraph 6.120, and
shown in data on order levels and evidence submitted by third parties) and so
the incentive to offer this service may have increased. In addition, Amazon
has benefitted financially from Coronavirus (COVID-19), with increased sales
and demand []. Its share price has increased and is up 36% year-over-year
(as at 11 June 2020), showing a clear expectation that the Coronavirus
(COVID-19) pandemic will ultimately advantage Amazon’s business.
Conclusion on Amazon food strategy
6.124 Evidence shows the current priority for F3 is to expand Amazon’s position in
online groceries globally, with a strong focus on [], and it is investing heavily
in its delivery capabilities in the US and UK. Despite this, interest in restaurant
delivery has continued for the past five years, is international, and extends
within the F3 team, as well as to senior executives at the highest level within
Amazon, including Jeff Bezos (CEO). The F3 team (as well as Jeff Bezos,
CEO) was supportive of Amazon [] following the closure of Amazon
Restaurants, and saw this as []. There is no evidence to suggest Amazon is
no longer interested in restaurant delivery or that it no longer expects it to be
an important area providing benefits such as differentiation in its offering,
flywheel effects for Prime, and enhanced logistical capabilities. This is further
supported by the recent trial by Amazon of entering restaurant delivery in
India.
Amazon Restaurants
6.125 This section discusses the launch, operation and closure of Amazon’s
restaurant delivery business, Amazon Restaurants, which was active in both
the UK and the US.
6.126 The Parties submit that Amazon has already tried and failed to operate an
online restaurant platform, which it chose to abandon only recently because it
realised it was unable to succeed. They submit that the Amazon Restaurants
145
experience was so []. They submit that Amazon Restaurants failed on all
three sides of the market: [].
6.127 A summary of the performance of Amazon Restaurants and its objectives are
set out in its ‘Operational Plans’ (OPs). Amazon explained that these are
generally yearly documents produced by multiple teams within a business
area []. [], and the business area will execute based on the OP1 meeting.
[] very senior executives are invited to OP meetings and discussions, and in
some cases are responsible for signing them off.
302
Amazon’s Finance team
also need to sign off on the figures included in the OP1 documents.
6.128 Amazon submitted that the Amazon Restaurant OP1 documents should be
treated with caution as they were produced by a team working hard to build a
business they believed in, and that, given the closure of Amazon Restaurants,
the team was [] the importance of restaurant food delivery services to
Amazon. Although we understand the team working on delivering Amazon
Restaurants might have been optimistic about performance, we note that OP1
documents must be signed off by senior executives (who typically would not
have the same incentive). We have not seen evidence of concerns raised on
the achievability of the OP1 documents or evidence indicating senior
executives did not share the vision contained in the documents. We therefore
consider these documents are a good source of evidence of Amazon’s
interest in and incentives to offer a restaurant platform, notwithstanding
Amazon Restaurants’ failure to achieve some of its targets prior to closure.
Launch (2015-2016)
6.129 Amazon Restaurants was created as a result of insights from Amazon Local,
a service that sourced daily deals for customers from local merchants in
various categories, including discounted vouchers for use in restaurants.
303
These insights showed that customers []. Amazon Restaurants was trialled
in Seattle, and then launched more widely in 2015, when Amazon started
building its own delivery service for restaurants and added Amazon
Restaurants to its Prime Now offering. Amazon Restaurants was launched in
the UK in 2016.
6.130 The Amazon Local OP1 dated November 2013 proposes an Amazon Local
delivery service that would provide delivery of items to customers in sixty-
minutes or less starting with restaurant orders but that ‘over time we plan to
302
For example, an Amazon internal email []. Amazon submitted lists of invitees to different types of OP
meetings [].
303
During 2014-2015 Amazon launched three new businesses following insight from Amazon Local, which
related to restaurants, travel and ticketing. In 2015 Amazon decided to discontinue its daily deals business. It
shut down its travel business in October 2015 and its ticketing business in 2018.
146
experiment with adding additional selection, []. This document notes the
benefits to and from Amazon’s wider business of adding this service, stating
that ‘[]’.
6.131 Subsequent documents from around launch in 2015 state Amazon saw [].
The document states Amazon will focus on [], including [].
6.132 Amazon submitted that the evidentiary value of these documents was very
low because of their age and the fact that the imagined developments never
materialised. Although in the context of a dynamic market these documents
are old, we consider, in combination with other evidence, that they show that
Amazon has a long-standing interest in restaurant delivery and has held an
aspiration to include this as a service within the Amazon ecosystem
internationally since around 2015.
Operation (2016-2018)
6.133 The Amazon Restaurants OP1 documents for this period provide evidence of
Amazon’s expectations for its restaurant delivery business and show the
online restaurant platform business has strategic value for the wider Amazon
business. These include that restaurant delivery will be an important part of a
wider food offering for consumers, [], and that the technology used for the
fast delivery of restaurant food will benefit the wider Amazon business as it
can be rolled out to other areas, [].
(a) The Amazon Restaurants OP1 2017 states that one of the five key
learnings from its first year of [].
(b) The Amazon Restaurants OP1 2018 reiterates the impact of offering
restaurant delivery in the US on Prime, [].
(c) The appendix to the 2018 OP1 document states that because of shifts in
consumer spending towards prepared food and the growth of takeaway
and delivery, Amazon believes ‘[]’. The document sets out [].
(d) Amazon Restaurants’ final OP1 (dated August 2018) discusses how the
technology used within Amazon Restaurants ([])
304
could be rolled out
to other business areas as it ‘[]’. The inclusion of restaurant delivery as
part of a wider food offering is further discussed, [].
304
[]
147
6.134 The Amazon Restaurant OP documents discuss some of the issues it is
facing globally with restaurant delivery, which continue from launch through
the operation of the UK and US businesses. For example:
(a) The 2017 document lists problems such as [].
(b) Evidence from the US OP1 documents shows Amazon Restaurants [].
For example, in the 2018 document, it achieved a cost per delivery of
$[] compared to the OP2 target of $[], and $[] million of gross
revenue, compared to a $[] million target.’ The document comments
that ‘[]’. It mentions ‘[]’.
(c) The 2018 OP1 for the UK business states [].
6.135 These documents demonstrate that one of the key issues for Amazon
Restaurants, both in the US and UK, was the []. The Parties also submitted
that this was a reason for its failure, and that the logistics for restaurant
delivery (generally point-to-point delivery with drivers paid per delivery) are
very different to that of its other businesses (generally point-to-multipoint with
drivers paid per block of time).
6.136 Amazon launched restaurant delivery using its [] logistics solution. In the
UK [].
305
This solution, []. Amazon looked to address this by [].
Although this was a significant improvement, it still indicates unsustainable
delivery costs in the UK – one Amazon Restaurants document indicates [].
As discussed in paragraph 6.146 below, Amazon was more successful in
bringing down delivery costs in the US.
6.137 Amazon submitted that in an attempt to resolve the challenges it was facing
(before the decision to shut the Amazon Restaurants UK business was
taken), it started ‘[]’. We note that the Amazon Restaurants UK OP1 for
2019 []. Shortly after this document was produced, the UK business was
closed.
6.138 Amazon submitted that the Amazon Restaurants OP documents described
above are old and that they show expectations for the business that were
never met, and so have a low evidentiary value. We disagree with these
statements. First, some of these documents are from 2018 and produced less
than a year before the contemplation of the Transaction, which we would
normally consider (absent specific and evidenced reasons to the contrary)
provide useful insight into the strategy and incentives of a company prior to a
305
Amazon Flex is an Amazon organisation that manages flexible drivers for Amazon, who are offered jobs
containing blocks of time between one and six hours.
148
merger. We further note that many of the positions set out in these
documents, in particular in relation to the benefits of offering an online
restaurant platform, are consistent with those set out in Amazon’s more recent
documents (as well as its older documents from the launch of Amazon
Restaurants). Second, although Amazon Restaurants did not achieve the
success hoped, important aspects of its performance appear to have been
positive. For example, the Amazon Restaurants business created technology
that Amazon is now using in its wider business (which was originally regarded
as one of the important benefits of developing and offering restaurant
delivery); the [] technology, originally developed for the Amazon
Restaurants business, has been []. Third, understanding Amazon’s
rationale for operating restaurant delivery in the first place, and the weight
placed on differentiating Prime Now and providing an important benefit for
Prime customers, indicates its incentives for originally entering this market,
which may also be relevant to any future entry. We note, in this regard, that
many of the factors considered at that time also feature in more recent
internal documents relating to Amazon’s broader strategy, such as Amazon’s
entry into restaurant delivery in India.
UK Closure (Late 2018)
6.139 As mentioned in paragraph 6.134, Amazon’s internal documents show it
experiencing a number of issues with its UK Restaurants business, including
that it struggled to []. These factors led the Amazon Restaurants team to
recommend closure in the UK in [], which was actioned in November 2018.
The documents show:
(a) Amazon Restaurants struggled to find the right [] and [], and only
implemented [] in the UK in April 2018. One document notes ‘[]’.
Even once Amazon implemented [], the lowest cost per delivery
Amazon Restaurants UK was able to achieve was still £[]. By
comparison, the Parties submitted that Deliveroo’s current cost per
delivery is approximately £[].
(b) Amazon Restaurants UK used only a [], meaning its logistics network
was less effective than in the US.
(c) The service was not []or available to non-Prime members (either in the
UK or US). Amazon spent [] amount on marketing the service in the
149
UK
306
and the documents suggest the product was not []. For example,
the 2017 Amazon Restaurants OP1 stated it has ‘[]’.
(d) Amazon Restaurants UK did not expand outside of London and did not
grow its restaurant selection in line with Uber Eats, which launched in the
UK at a similar time.
307
Know-your-customer (KYC) regulations led to
Amazon Restaurants UK []. Amazon’s submission states that ‘[]’. The
2018 OP1 document warns of this issue, stating that in order to complete
these checks Amazon Restaurants ‘[]’. One restaurant that was part of
the Amazon Restaurants platform told us that it had difficulty with both the
platform and Amazon’s processes.
(e) It failed to compete effectively with [] and other competitors in []. [],
with one Amazon document stating it encountered ‘[]’. Amazon
submitted that it did not sign [] with any restaurant in the UK, as the
business was not sufficiently attractive to restaurants for them to [].
6.140 These challenges were reflected in [] in an Amazon customer survey, [].
[], Amazon decided to recommend that the UK business was shut down
[].
6.141 We consider that the evidence shows that Amazon appears to have made
some strategic and operational errors or misjudgements in the way it
implemented Amazon Restaurants in the UK and may have underestimated
the difficulties in building such a business. Although it took some remedial
action and staff working within Amazon Restaurants were confident in
improving the business, it was not sufficient to turn the struggling business
around to the satisfaction of Amazon’s senior executives.
(a) First, we observe that successful entry as a logistics-enabled online
restaurant platform appears to require substantial investment and
willingness and ability to incur losses in the early years in order to invest
in the growth of the business to build density of customers and
restaurants.
308
It does not appear []. The absolute level of costs
incurred and losses made were []. Amazon Restaurants had cumulative
[] in the UK from 2016 to 2018 of $[] million, with total costs incurred
of $[] million. As a comparison, Deliveroo incurred total operating
losses of £161 million ($220 million)
309
from January 2014 to December
306
[]
307
Uber Eats told the CMA that Amazon Restaurants launched at a similar time to it with three times more
restaurants available but that Uber Eats accelerated and overtook Amazon Restaurants in terms of selection.
308
As acknowledged by Amazon in an internal email which states restaurant delivery is a [].
309
Using average £/$ exchange rates of 1.6455 (2014), 1.5278 (2015) and 1.3502 (2016).
150
2016, following its foundation in August 2013, and costs of £270 million
($365 million) in 2016 alone.
310
Amazon submitted that the [] in the UK
do not include [] in the US that would have benefited the UK business,
such as []. The [] by Amazon Restaurants in the US were [] than
those in the UK
311
but we note that the figures for the total amount [] by
Amazon Restaurants during operation in the UK and the US are [] the
amount Deliveroo lost in the EU in one year alone.
(b) Second, Amazon did not [] a large customer base. The service was []
and available to existing Prime customers in a certain geographic area.
Amazon had significantly fewer Prime customers in 2016 ([] million) in
the UK than it does now, with Prime membership [] in the last three
years (to [] million). Amazon spent [] on marketing the service in the
UK, which was likely to have been an important factor in why [] Prime
customers used Amazon Restaurants. A breakdown of Amazon
Restaurants’ UK costs shows that in the three months of 2016 in which it
operated, Amazon Restaurants spent $[]k on marketing and business
development, [] to $[]k in 2017 and $[]k in 2018. As a comparison,
Deliveroo spent £[] million on marketing in the UK alone in 2019, [].
Amazon Restaurants was also only available to customers in certain
London areas.
(c) Third, when asked what Amazon learnt from operating Amazon
Restaurants, Amazon told us that [].
(d) Fourth, internal documents show evidence of Amazon considering that it
did not [] in place or [] in the Amazon Restaurant business. Amazon
Restaurants employed under [] staff during its operation in the UK,
although the 2019 UK OP1 document shows a fixed headcount of only
[]. Worldwide, Amazon Restaurants had a total headcount of just over
[]. This is in comparison to the estimated staff needed under Amazon’s
[]
312
it produced when considering whether to [], which included a
total of [] people made up of [] engineers, [] in marketing, [] in
account management and [] support and tech staff. One Amazon
internal email written by [] (VP, Corporate Development) states that
Amazon Restaurants’ offering in 2018 is ‘[]’. A different email written by
[] (VP Corporate Development) states that he thinks ‘[]’, to which
[], the SVP in charge of Amazon Restaurants, [].
310
Roofoods Ltd Annual Accounts Year Ended 31 December 2014, 2015, 2016 (see Companies House).
311
Amazon Restaurants lost $[] million with total costs incurred of $[] million.
312
Although we consider this analysis was simplistic, we consider the staff figures included are helpful for context
for considering the staffing figures of Amazon Restaurants.
151
6.142 Amazon submitted that the evidence described above supports its submission
that it has no compelling incentive to re-enter this segment, as if it did, it would
have [] and continued to operate in the market. Amazon provided evidence
showing it invested [] in its US offering, which would indicate it did have an
incentive to operate restaurant delivery as a Prime benefit, as part of its
overall food strategy, but that it was not []. It was instead focussed on [].
We do not consider that Amazon’s decision to exit the UK means that it was
no longer interested in this market – indeed, just months after the closure of
the UK business it began due diligence on potentially [], showing clear
interest in the UK restaurant platform market.
6.143 We received some evidence that showed other suppliers active in the sector
saw Amazon Restaurants as a learning experience by Amazon and that they
would expect it to try again.
313
Amazon was seen as a credible potential
entrant notwithstanding its previous departure from the market. Although
these views are helpful in our assessment, we note that these parties would
have limited insight into Amazon and its strategy and so gave such evidence
less weight in our assessment. Their views that Amazon would look to ‘test
and learn’ were however consistent with the approach that [].
US Closure (Mid 2019)
6.144 Amazon continued to operate its restaurant delivery business in the US after
shutting down the UK business, but submitted that due to [], it closed
Amazon Restaurants US in June 2019.
6.145 Amazon invested [] in its US offering than the UK one, and it was available
in 20 US cities. It incurred total costs between 2016-2019 of $[] million, with
[] of $[] million. The figures for the total costs incurred by Amazon
Restaurants during operation were [] the costs incurred by Deliveroo in a
similar timeframe (see paragraph 6.141(a)), and total [] the amount
Deliveroo lost in the EU in one year alone (Deliveroo had an operating loss in
2018 of £257 million ($342 million). Amazon’s total marketing spend in the US
was [] that spent by Deliveroo on marketing in the UK in one year alone (in
2019, Deliveroo spent £[] million ($79 million).
314
6.146 The 2019 OP document shows some successes for the US business – it [].
A 2019 planning document for Amazon Restaurants expects []. There is no
indication in these documents, written during 2018, of plans to wind down or
close the US business, or any indication that the closure of the UK business
313
This includes a Deliveroo internal email where [] and [].
314
Using a 2019 average exchange rate £/$ of 1.2772.
152
has impacted the US business, despite these documents being reviewed by
senior management.
6.147 In late February 2019, [] (SVP, Physical Stores and F3, including Amazon
Restaurants) states in an email to [] (VP Corporate Development) that he
has []. A communications plan for the closure of Amazon Restaurants US
produced for Amazon senior management in 2019 discusses how, although
Amazon Restaurants is to be closed, it is possible that Amazon would re-enter
in the future. The document states []. This document also mentions that
technology used by Amazon Restaurants will be used elsewhere in Amazon,
stating that ‘[]’.
6.148 Both extracts from the communication plan document indicate the potential for
Amazon to re-enter this area in the future and to use the knowledge and
technology gained from Amazon Restaurants in the future across different
countries. Amazon submitted that this document was an internal
communication plan intended to ‘[]’ and any suggestion of re-entry by
Amazon in this document is speculative. We note that Amazon is now using
the [] developed for Amazon Restaurants in its US grocery business and
has started piloting a restaurant delivery service in India following the transfer
of the technology, showing that the comments in the document were more
than platitudes. This also shows that Amazon plans to continue using and
developing aspects of its restaurant delivery business that could in the future
be used again.
6.149 Evidence shows the decision to close Amazon Restaurants was not [] and
the senior staff involved agreed that it did not rule out future entry. An email
from 25 March 2019 shows [] (VP, F3) stating her [].
315
The reasons
given are that it is a []. On 27 March 2019, [] replies to state he is ‘[]’
but that he doesn’t see closing the business as a ‘one-way door decision’ and
believes Amazon could re-enter in the future []. The final decision on
closure is left for Jeff Bezos to make, with [] (SVP, Physical Stores and F3,
including Amazon Restaurants) chasing a response from him on 2 April 2019.
Amazon submitted this decision was made during a call attended by these
individuals on [] 2019, and provided the calendar invite to support this
position.
6.150 We consider the date of this call to be relevant, as there appears to be a delay
between the point at which a decision on the future of Amazon Restaurants
US was requested and the timing of the call when the decision was made. As
noted above, that decision was made during a call on [] 2019, which took
315
This is an email chain between [] (SVP of Physical Stores and F3), [] (VP, F3), Jeff Bezos (CEO), []
(CEO of Worldwide Consumer) and [] (CFO).
153
place [] after the meeting on [] 2019 where Jeff Bezos decided to invest
in Deliveroo. Amazon has submitted that the decisions to invest in Deliveroo
and to close Amazon Restaurants were taken separately. Given the lack of
evidence provided by Amazon on the reason for its decision to close Amazon
Restaurants (consistent with Amazon’s broader submission, as noted in
paragraph 6.78, that it does not []), we note that the only evidence that is
available shows the decision to close Amazon’s restaurant delivery business
was taken [] after deciding to invest in Deliveroo, and both decisions were
taken by the same person, Jeff Bezos. The CMA has repeatedly provided
Amazon with opportunities to clarify the rationale for these decisions,
including through multiple requests for information and requesting an
appropriate attendee at the main party hearing who could discuss these
decisions, but Amazon has declined to do so. Given this position, and the
evidence discussed later from paragraph 6.154 around Amazon’s decision to
invest in Deliveroo,
316
we find it likely that the two decisions were linked. We
therefore consider that the decision to close Amazon Restaurants US may
have been influenced by Amazon having an alternative route to enter
restaurant delivery internationally.
Conclusion on Amazon Restaurants
6.151 Amazon’s previously unsuccessful attempt in restaurant delivery does not
detract from its continued global interest in food delivery (see paragraph 6.120
onwards) and its incentives to offer such a service, as indicated by Amazon
India’s entry in this space. These efforts show an intention by Amazon to be
present in this space. Amazon’s documents emphasise the importance and
benefits to Amazon of offering restaurant delivery, including being in a large
and growing market, [], an additional benefit to power Amazon’s ‘flywheel’,
and an area where useful logistical technology can be developed and later
expanded to other products []. Amazon did not manage to execute this
vision in the UK, where it []. It also struggled in the US and did not appear
to be willing to [].
Amazon’s interest in Deliveroo
6.152 Amazon first responded to Deliveroo’s bankers indicating its interest in
Deliveroo in []. It initially considered [], after which it considered making
an investment []. It decided to make a minority investment (the Transaction)
in [] 2019.
316
In particular an Amazon internal email where Jeff Bezos is reported [].
154
6.153 The Parties told the CMA that the decision to close Amazon Restaurants and
the decision to invest in Deliveroo were taken separately. Amazon
Restaurants closed in the UK in November 2018 (prior to interest in Deliveroo)
and in the US in June 2019, with the decision on closure made in [] 2019.
6.154 As discussed in paragraph 6.150, the timing of the decisions to close Amazon
Restaurants in the US and invest in Deliveroo overlap, and the ultimate
decision maker was the same in both cases (Jeff Bezos).
317
There is
additional evidence of a connection between the decisions in an Amazon
email between the corporate development team in mid-February 2019, in
which [] (VP, Corporate Development) reports on his catch up with []
(SVP, Business Development) and states ‘[]’.
6.155 The interest in Deliveroo appears to be []. For example:
(a) In Jeff Bezos’ email to [] (VP, Corporate Development) on
16 November 2018, he states that Deliveroo’s banker ‘[]’. An email from
[] (VP, Corporate Development) on 19 November 2018 informs a small
group of staff (including []) that corporate development were instructed
to look at Deliveroo [], and another email thread from the same day
from [] (VP, Corporate Development) circulated within the corporate
development team states that Jeff Bezos ‘[]’.
(b) [], an SVP in Amazon, states that he is in agreement with [], another
SVP, that []. In this thread, [] (VP, Corporate Development)
expresses [].
(c) [] (VP, Corporate Development) states that he [].
(d) In an internal email discussion between [] (VP, Corporate
Development) and Jeff Bezos, []suggests Amazon [].
6.156 Amazon’s rationale [].
6.157 Amazon’s briefing regarding [] of Deliveroo states ‘… an [] of Deliveroo
could be interesting for Amazon for the following reasons… []’. Amazon
submitted that these are uninformed hypotheses made by [], although we
note that factors such as flywheel benefits, low cost delivery and it being an
attractive and growing market are similar to those considered by the F3 and
Amazon India teams who work or worked in restaurant delivery businesses.
317
[]
155
Amazon build vs buy analysis
6.158 As part of corporate development’s due diligence on whether to [] invest in
Deliveroo, Amazon produced a ‘build vs buy’ analysis. This compared the
NPV of building an equivalent footprint to Deliveroo’s UK/EU business in []
years (by []) to the returns from that business over the same period. This
showed that ‘build’ had a [] whilst ‘buy’ produced a [] (excluding the initial
purchase price) for the years []. The text summarising the analysis in the
report produced by the corporate development team states:
‘[]’.
6.159 Amazon submitted that this analysis showed that ‘building a UK online
restaurant food delivery service from scratch would be []’ and that ‘Amazon
believed that de novo entry into the UK and EU online restaurant food delivery
would incur a [] over []’ and ‘[]’.
6.160 It is clear from the wording in the report that this is not an accurate
representation of what the analysis showed. The analysis looked at a set time
period ([]) and so does not include all relevant costs (or benefits) in order to
conclude the NPV of re-entry by building a restaurant platform would be ‘[]’.
Instead, the NPV from building an equivalent to Deliveroo in the UK/EU in []
in that period. The differential between these two was then [] which is not
included in the [] year ‘buy’ model. Once an allowance is made for the [],
the NPV for ‘build’ is substantially higher than the NPV associated with ‘buy’
(with a differential of £[]).
6.161 Furthermore, as the assumption of the ‘build’ model is that Amazon would
have built an equivalent business to Deliveroo at the end of the []
timeframe, it follows the cashflows in both the ‘build’ and the ‘buy’ model after
[] should be similar. As the buy model’s total NPV is £[] (from which the
purchase price is derived) £[] of its NPV must be realised after [] (as the
NPV for [] was estimated at £[]). Assuming a similar amount arises in the
‘build’ model after [] would mean it would have a [] (of £[]).
318
The
briefing paper [] does not reach an explicit conclusion about the
implications of the NPV analysis for the relative attractiveness of ‘build’ versus
‘buy’. However, when analysed and presented on a consistent basis, we
consider that the analysis implies that the NPV of the ‘build’ option is
materially higher than the NPV for the ‘buy’ option. Further, the analysis does
318
With this allowance, the analysis showed a differential of £1.176 billion between the two models compared to
a purchase price of £2 billion (UK/EU Deliveroo) which means that the ‘build’ model is in effect £824 million better
from a purely financial perspective than the ‘buy’ model.
156
not in our view support Amazon’s position that the NPV of re-entry by building
an online restaurant platform would be ‘[]’.
6.162 We would finally note that the ‘build’ model is relatively simplistic and the
report notes a number of non-financial risks that could impact on the analysis.
As such, we do not consider it to have strong evidential value in showing that
Amazon has an intention to re-enter the supply of online restaurant platforms
in the UK organically.
Amazon’s decision to invest in Deliveroo [
]
6.163 At a meeting held on 12 March 2019, Amazon decided [] pursue a minority
investment. Amazon told the CMA it decided [] Deliveroo because [].
319
Amazon has, however, not been able to provide any documentation of this
meeting, or any contemporaneous documents that clearly set out the reasons
for Amazon’s [] in approach.
6.164 The rationale for the minority investment in an Amazon internal briefing paper
starts as in the document referred to at paragraph 6.156 above (stating that
Amazon needs to offer a []) but then continues instead to state ‘[]’.
Amazon told the CMA that the rationale text had been drafted [], and had
been copied into a subsequent email relating to the investment decision,
without amendment, and was not relevant to the investment decision. This
does not appear to be credible though, as parts of the text had been updated
to take account of the change [] to invest (as demonstrated by the specific
reference to ‘the investment’ and the related rights in the final sentence).
Additionally, the document in question was a short email sent to Jeff Bezos
(CEO) and Brian Olsavsky (CFO) seeking approval for a significant
investment, and may therefore reasonably be assumed to have been carefully
drafted and reviewed.
6.165 Amazon submits that the clearest evidence that it would not re-enter
restaurant delivery in the UK is its own revealed preference in that it decided
[]. We consider that by investing significantly in Deliveroo that Amazon has
clearly revealed an interest in restaurant delivery in the UK and that it is still
interested in participating in this market despite having exited in the past.
Amazon told us that it saw the investment as providing ‘a little bit of
optionality’
320
around investing further or acquiring Deliveroo in the future,
319
We note Amazon’s valuation of Deliveroo forecast a cash burn of approximately $[] from 2019 to 2022. If
Amazon [] Deliveroo, Amazon would have been the sole source of this cash, whether or not Deliveroo is
ultimately successful.
320
The CMA recognises the argument in Deliveroo’s response to the Revised Provisional Findings (10 July 2020
at paragraph 6.3) that the Transaction does not provide Amazon with any ‘option’ to acquire Deliveroo. This
157
although stated it [], and that any such further investment would be subject
to CMA review. In addition, we consider that there were multiple options for
Amazon to re-enter the UK (as discussed further from paragraph 6.168) and
so acquiring Deliveroo was not the only way it could re-enter.
Conclusion on interest in Deliveroo
6.166 Amazon’s CEO, Jeff Bezos, has shown clear interest in Deliveroo, and
evidence shows []. Even though Amazon [], evidence indicates that it
sees restaurant delivery as ‘strategically aligned’, being a large and growing
market and was seen as [] while providing flywheel benefits.
Conclusion on Amazon’s incentives and intention
6.167 On the basis of the evidence set out above, we believe that Amazon has a
strong and continued interest in online restaurant platforms and an incentive
to offer this to Prime customers in order to differentiate its offering, realise
flywheel benefits, and develop useful logistical capabilities that would be
deployed elsewhere in its business. Evidence shows that the UK is a large
and growing market for online restaurant platforms, and the UK is an
important and attractive market to Amazon as a result of high levels of Prime
membership. Since the closure of Amazon Restaurants, we have observed in
Amazon’s internal documents continued global interest in online restaurant
food delivery. [].
CMA assessment: Amazon’s ability to re-enter
6.168 As well as considering whether Amazon has the incentive to re-enter the
supply of restaurant platforms in the UK, we also need to consider whether it
had the ability to do so absent the Transaction, ie it has one or more routes to
enter the market.
6.169 There are a number of possible routes that Amazon could take to re-enter the
supply of online restaurant platforms. We believe there are three main ones;
(i) building its own offering, (ii) acquiring or investing in an existing restaurant
platform or adjacent business; and (iii) partnering with an existing restaurant
platform or adjacent business. These three routes are discussed below. For
the purposes of establishing the counterfactual in this case, consistent with
the CMA’s established framework for analysis
321
(including as set out at
statement quotes Amazon’s own characterisation of the Transaction during the main party hearing and is not
merely a conclusion that the CMA reached based on its other findings.
321
MAGs, paragraph 4.3.6. The application of this can be seen in the CMA;s recent investigation into PayPal’s
acquisition of iZettle, PayPal/iZettle, paragraph 7.35.
158
paragraph 6.3 the need to avoid spurious claims to accuracy of prediction or
foresight), we need only consider the viability of these options with a view to
determining whether the most likely scenario is that Amazon would enter via
one or other of these routes. We do not need to specify which route Amazon
would take.
6.170 The Parties submitted there are high barriers to entry in the supply of online
restaurant platforms in the UK, and that Amazon is not particularly ‘well-
placed’ to overcome these. Barriers to entry are discussed in more detail from
paragraph 6.197 but mainly consist of: (i) commercial barriers to recruit
restaurants, couriers and consumers, and (ii) technological barriers in terms of
developing a platform that interfaces with all three sides of the market and
that powers a point-to-point delivery service.
322
The Parties submitted the
previous failure of Amazon Restaurants shows it does not have the ability to
re-enter, and that the barriers to entry are higher than when Amazon
previously tried to enter (see discussion of barriers from paragraph 6.197).
Build
6.171 One possible route of entry for Amazon to supply an online restaurant
platform in the UK would be to build its own business organically.
6.172 Amazon previously attempted to operate in the UK market with Amazon
Restaurants, as discussed in detail above from paragraph 6.125. In particular:
(a) Amazon only incurred total costs of $[] million in Amazon Restaurants
in the UK over the total time it operated it (plus costs incurred in the US,
such as technology development), and the service was not [].
(b) Amazon used technology and a logistics [], although later did invest in
[]. Evidence shows [].
(c) Its offering received poor customer feedback [].
6.173 The evidence indicates a poorly executed attempt by Amazon at offering an
online restaurant platform in the UK during 2016-2018, and its overall Amazon
Restaurants business is described in one internal email as a ‘[]’ and in
another as []. Amazon’s [] to improve its online restaurant delivery
business and decision to instead close it down in the US, and the nearly
contemporaneous decision to invest nearly $[] million in Deliveroo could
indicate Amazon would prefer to buy rather than build a restaurant delivery
offering. This is supported by an Amazon internal email from February 2019
322
Parties’ Initial Submission, 24 December 2019, paragraphs 4.10–4.14 and Annex 1.
159
where it is remarked that []. On the other hand, Amazon may have decided
that, given the various ways its restaurant proposition was failing to perform
and the risk of damaging relationships with Prime customers from offering a
poor quality service, that it was easier to withdraw Amazon Restaurants from
the market, invest in addressing the shortcomings that had been identified,
and to relaunch.
6.174 Having requested evidence on this point, we did not observe any material
indicating current plans or discussions in relation to Amazon re-launching
Amazon Restaurants in the UK (or US) through building such a business
organically. We note that the time between closing Amazon Restaurants in
the UK (August 2018) and exploring [] Deliveroo ([]) is very short.
323
Indeed one email discussing the F3 team’s view on [] Deliveroo states that
‘[]’.
6.175 As discussed above at paragraph 6.158, we assessed Amazon’s ‘build vs
buy’ analysis for the proposed acquisition of Deliveroo and found that
although this shows a [] cashflow for ‘build’ for [], the paper suggests a
[] overall NPV of building an online restaurant platform, which is [] the
NPV of buying Deliveroo once the purchase price is taken into account. The
model is high-level and notes non-financial risks associated with building
which are not considered further, but we consider the analysis does not
support Amazon’s submission that ‘build’ would be unattractive (although nor
do we consider it persuasive evidence showing that Amazon would build an
online restaurant platform).
6.176 We note that Amazon has decided to build its own online restaurant delivery
service in India. The [] indicates that Amazon Restaurants allowed Amazon
to learn and develop technology that is useful in operating this type of
business in the future, and that aspects of the technology are transferrable
between different countries. It appears to have taken Amazon [] to launch a
trial of its service in India []. Third parties have told us that the technology
powering their platform and delivery network is generally transferable between
countries and that you do not need country-specific data to launch. Amazon
would still need to overcome the other main barrier mentioned, of attracting
customers, restaurants and couriers to its offering in the UK, even if [],
although as discussed in paragraph 6.197 we consider Amazon has a
significant advantage at attracting customers as a result of Prime.
323
In addition, as discussed, Amazon decided to invest in Deliveroo [].
160
Conclusion on build
6.177 We have not observed evidence of Amazon assessing how it might build an
online restaurant platform in the UK, although we note that it is currently
planning to do this in India []. Amazon also told us that if such a service was
successful, it is []. We consider this shows Amazon’s ability and intention to
transfer learning across countries. Given the global nature of the sector, as
well as Amazon’s international strategy in offering this service as part of
Prime, we consider this evidence a relevant factor in our assessment of
Amazon’s global strategy in online restaurant platforms, including in the UK.
Amazon Restaurants in the UK appeared to fail due to [], and the difficulty
of overcoming the technological barrier of point-to-point logistics. As noted in
paragraph 6.158 above, we consider that the ‘build vs buy’ analysis
conducted by Amazon’s corporate development team shows a [] NPV for
‘build’ than it does for buying Deliveroo including the purchase price. This
analysis is high-level and so we do not consider this evidence persuasive in
showing that Amazon would build its own service. We consider, however, that
the evidence overall shows that building an online restaurant platform is a
plausible option for Amazon in order to enter the UK market.
Acquire or invest
6.178 A second option available to Amazon as a route to entry into the supply of
online restaurant platforms in the UK absent its investment in Deliveroo is
through acquiring or investing in an alternative business. Amazon has
significant financial resources available to it if it were to choose to invest or
acquire in this area.
6.179 As discussed above at paragraph 6.155, interest in [] Deliveroo came from
Jeff Bezos, the CEO and founder of Amazon, and the decision to invest was
also made by him. Amazon internal emails from November 2018 suggest an
overall continued interest in the restaurant delivery sector [], as []
comments that ‘[]’. There are other references where it appears Jeff Bezos
is particularly interested in [].
324
6.180 The Parties submitted Amazon [], and that Amazon [], and there is
accordingly no evidence of this.
325
6.181 As we would expect, given Amazon’s internal [] (as discussed at
paragraph 6.78) where it does not produce significant documentation
324
For example, an Amazon internal email where the [] an article discussing high levels of investment in cloud
kitchens [].
325
Parties’ Initial Submission, 24 December 2019.
161
concerning its discussion and decisions, most of the evidence we have seen
from Amazon relating to potential acquisitions or investments is high-level
consideration and initial scoping discussions by Amazon senior executives
rather than detailed consideration of the relevant business, and most relates
to businesses that operate outside the UK. We note that Amazon does not
appear to have [] discussing [] investment in Deliveroo, rather there was
[]. We would not therefore expect to find extensive documentation of
potential targets of Amazon.
6.182 We observed some evidence of Amazon’s interest []. This includes an
email where a member of the corporate development team states that ‘[]’
and another where a different member of the corporate development team
states he believes []. At this stage, we cannot rule out the possibility that
these transactions would raise competition concerns and as such we have not
considered either of these as potential routes for Amazon’s re-entry in the
counterfactual, but do consider the documents show a preparedness by
Amazon to acquire in order to enter the UK market.
6.183 We observed evidence of Amazon considering how to develop its restaurant
delivery business in the US. This includes:
(a) One email between [] (CEO) and [] (VP, Corporate Development)
from November 2018, where the corporate development team suggests:
‘[]’ to which Jeff Bezos replies: ‘[]’. Amazon submitted that this email
was just comparing other players to Deliveroo as informal benchmarking
but we disagree with this interpretation given the email clearly states that
these companies are considered in the context of what acquisition targets
exist for Amazon.
(b) In one email Amazon asks Deliveroo if Amazon would need to [] US in
order to compete there, with Deliveroo responding ‘[]’. This and the
above email indicate that Jeff Bezos believes Amazon should [],
326
and
if it did invest [].
(c) Other Amazon internal emails show senior executive interest in meeting
with [] to discuss a Prime Now partnership (discussed at
paragraph 6.190). [] is active in North America, and recently expanded
to [].
(d) [].
326
This idea is also supported in other internal documents, such as [], where a member of the corporate
development team states ‘[]’.
162
6.184 We also observed evidence of Amazon considering restaurant delivery
internationally outside of the UK and US (although none of these were
undertaken). This includes:
(a) An email chain from Amazon India in June 2019 shows its interest in [].
An Amazon India document from January 2019 states ‘[]’, which
appears to be why Amazon is unable to invest in these businesses.
(b) Another internal email chain between Amazon India and corporate
development from January 2019 discusses a potential investment in
[].
327
[].
(c) Detailed consideration of [].
(d) Other emails consider investment in businesses active in restaurant
delivery or adjacent to this, including [].
328
A service such as this could
help Amazon offer a restaurant delivery service by solving technological
barriers, with Amazon noting ‘[]’.
6.185 When in discussions with Deliveroo, Amazon told Deliveroo, ‘[]’. Amazon
submitted that this comment was made in the context of commercial
negotiations with the aim of generating commercial leverage and cannot be
treated as providing a reliable picture of Amazon’s commercial strategy. We
consider there were other ways Amazon could have generated leverage, that
Amazon would not have said this if it was not credible, and that the wording
used clearly indicates both that Amazon considered restaurant delivery
important and that it believes it has options on how it could participate in
restaurant delivery.
6.186 We consider Amazon acquiring an overseas online restaurant platform would
be unlikely to lead to Amazon entering the UK market in the very near term as
it would take time to understand and integrate the business. As described
above, we believe that the evidence shows Amazon has a clear incentive to
offer this service in the UK and that if it looked to enter through acquisition, it
would expand that service to the UK in the short to medium-term (ie within the
next five years, see discussion of timing at paragraph 6.199).
6.187 It is unclear what impact Coronavirus (COVID-19) might have on Amazon’s
ability to acquire or invest in a suitable business to facilitate entry into the
supply of online restaurant platforms. As stated in paragraph 6.123, the
pandemic has, and is expected to continue have, a positive impact on
327
[]
328
The focus for any partnership following investment by Amazon in these businesses was in ecommerce,
logistics and payments. These companies are considered in Amazon internal emails.
163
Amazon’s business. It is also possible that online restaurant platform
businesses, many of whom are loss-making, may have increased interest in
obtaining investment from a large investor such as Amazon. These factors
suggest Coronavirus (COVID-19) may make it less difficult for Amazon to re-
enter via an acquisition or investment. Overall, we consider there is
insufficient certainty around the impact of Coronavirus (COVID-19) to
understand whether this would enhance or delay acquisitions or investments
in these businesses.
Partner
6.188 A further option would be for Amazon to re-enter the supply of online
restaurant platforms through partnering with a business that either offers this
or would enable Amazon to offer this by addressing one or more of the
barriers to entry that Amazon faces. One of the key gaps submitted by
Amazon is its lack of a point-to-point logistics system. Potential partners to
address this could include last-mile delivery specialists, such as logistics
companies who operate point-to-point networks, as well as other types of
businesses who operate last-mile networks, such as ride-hailing businesses.
Uber Eats told us that an advantage it had when establishing was the logistics
technology it had developed for its main ride-hailing business and it was able
to build its platform from this. Partnering could help speed up the time taken to
enter the supply of online restaurant platforms in the UK as the parties could
combine resources and expertise, particularly if the partnering business had
experience in the market overseas. DoorDash, for example, told us that it
aims to enter new markets as quickly as possible, with it generally taking
between several months to one year.
6.189 Partnering appears to be an approach that Amazon takes in some business
areas. For example, it considered partnering in order to facilitate its entry [].
It also partners with grocery providers, for example Morrisons, to provide
groceries for its offering, as well as using their stores as fulfilment centres to
increase efficiency of delivery. One internal Amazon email we reviewed
suggests a partnership with a food delivery business [].
6.190 As mentioned above, food delivery company []. Amazon submitted it does
not control the opportunities that are presented to it, and it is clear from an
Amazon internal document that this has not been taken forward. We observe
that there was some interest from senior executives, including [] (CEO,
Worldwide Consumer) and [] (VP, Corporate Development) in meeting with
them to see if there was anything to learn, even if this did not lead to further
164
developments.
329
[]. We note the timing of this exchange is shortly after the
CMA’s phase 1 decision on Amazon’s investment in Deliveroo, where the
CMA explicitly set out the possibility that Amazon could partner with a non-UK
player in order to re-enter the supply of online restaurant platforms in the UK
and as such we cannot exclude that negotiations could have been affected by
CMA proceedings.
Availability of possible partners
6.191 We sought evidence from a wide range of third parties in order to understand
their current position and capabilities, future plans to enter the UK, openness
to partnerships, and appetite to work with Amazon. This included international
restaurant platforms
330
as well as UK-based point-to-point logistics specialists
such as Stuart and Gophr.
6.192 We consider that this is not an exhaustive list, and there are likely to be other
companies operating globally that could be potential partners or targets for
Amazon.
331
Amazon submitted that it gets many unsolicited approaches from
companies including those active in restaurant delivery and related markets
([]). Given the dynamic nature of food delivery and market growth, it is likely
that new businesses will emerge or develop in the near future, despite there
being uncertainty around who these might be.
6.193 Evidence shows that, as might be expected, companies are interested in
partnering with Amazon. One email from June 2019 mentioned that Amazon
might have ‘[]’, []. [] told us that it would be interested in partnering
with Amazon and believes they could have a mutually beneficial relationship
where it gains access to Amazon Prime customers and Amazon gains from
[]. Other third parties we spoke to, including [], expressed an interest in
partnering with Amazon. Difficulties in obtaining finance during the
Coronavirus (COVID-19) might increase interest in partnerships from affected
businesses who would benefit from the support Amazon could provide.
Benefits of acquiring, investing or partnering for Amazon’s ability to enter
6.194 The Parties submitted that barriers to entry in the UK market have increased
since Amazon previously attempted to enter. We tend to agree with this
329
Amazon response to [] also shows that Jeff Bezos is ‘[]’.
330
We contacted Takeaway.com (prior to its merger with Just Eat), Glovo, Delivery Hero, DoorDash, Postmates
(prior to its agreed merger with Uber), Grubhub (prior to its agreed merger with Just Eat Takeaway.com), Grab,
Meituan, DiDi, Rappi, Swiggy, Zomato, Bolt, and Wolt.
331
Third parties have also suggested Rico Logistics and Delivery Mates as point-to-point logistics operators who
could potentially power an online restaurant platform in the UK.
165
statement, as evidence shows the existing players have developed and
expanded their UK propositions in the past few years to improve their service
for consumers, restaurants and couriers. For example, []. Another example
is the expansion of delivery-only kitchens, which help restaurants deliver to
consumers that were previously out of reach.
332
6.195 The main barriers identified by the Parties were operational and technological
barriers in operating a three-sided platform, and commercial barriers in
attracting restaurants, couriers and consumers to the platform. We consider
that the benefit to Amazon of acquiring, investing in, or partnering with a third
party is that this approach would allow it to more easily and more quickly
overcome these barriers to entry in the supply of online restaurant platforms in
the UK.
6.196 []. Just Eat acquired Skip the Dishes in December 2016 and told us that it
believes this is the most successful acquisition it has made and has enabled
[]. Just Eat also partners with Stuart for delivery in the UK and told us that
[]. Just Eat believe it was key for them to be able to offer logistics in order to
win McDonalds as a customer.
6.197 As can be seen by the Just Eat example, partnering can allow a business to
develop capabilities and improve its offering in restaurant platform offering.
Such an arrangement (or acquiring or investing) could help Amazon
overcome the barriers to entry in supplying a restaurant platform and fix some
of the issues it encountered with Amazon Restaurants. This could include:
(a) Improving Amazon’s logistics technology used in restaurant
delivery: It is clear that other businesses have the necessary technology
to power an online restaurant platform, and they are able to use this to
offer delivery in a cost-effective way. For example, [] told us that it
delivers restaurant food in under 30-minutes and achieves 99% of its
SLAs []. A Deliveroo internal email indicates []. An Amazon internal
email states ‘[]’. Third parties told us that technology is transferable
between business areas and countries. For example, Uber Eats told us
that it was able to build its logistics technology on the back of the ride-
hailing business of Uber, which was an advantage when starting
restaurant delivery. [] told us the technology it uses is the same
regardless of what it is delivering (be it restaurant food, groceries, or retail
goods). Third parties also told us that technology can be deployed across
332
Deliveroo operates delivery-only kitchens (or ‘dark kitchens’) under the Deliveroo Editions concept.
166
different countries with minor country specific developments This could
facilitate a faster entry into the UK online restaurant platform market.
(b) Attracting restaurants to Amazon’s platform: Overseas restaurant
platforms may already have relationships with global or international
restaurant chains and would be interested in expanding that relationship
to include the UK. []. The relationship with McDonalds allowed Uber
Eats to build up scale quickly and provide jobs for its couriers.
Establishing [] restaurant relationships was seen as sufficient by
Deliveroo in order to enable Amazon to offer restaurant delivery in the US
as ‘[]’. Having [] restaurant relationships could quickly allow Amazon
to scale its offering, which would attract and retain couriers (or to offer
more jobs to its existing fleet of Amazon Flex drivers).
(c) Attracting and retaining couriers to Amazon’s platform: As described
above, businesses active in restaurant delivery or similar areas are likely
to be able to transfer their technology to the UK market, and can be used
to attract and compete for couriers through the ability to offer, for
example, ‘surge pricing’ at times of high demand. Uber Eats told us that it
benefited from the operational expertise from its ride-hailing business
when entering the UK restaurant delivery market.
6.198 An investment or partnership would be unlikely to address the barrier of
attracting consumers to Amazon’s restaurant platform. However, Amazon has
access to over [] million Prime subscription customers in the UK, a number
which has nearly [] compared to when it launched Amazon Restaurants. As
a result of having to pay a subscription fee and because of the wide range of
benefits available to subscribers (including Amazon Video, Amazon Music
and Twitch), these customers are likely to regularly engage with Amazon
products. This gives Amazon multiple avenues for engaging with customers
and marketing a new service to them and saving significant amounts on
marketing and advertising compared to other players. Prime customers would
be a unique audience of customers, which [] expressed interest in being
able to reach. [] told us that large restaurant chains want to reach increased
volumes of customers []. These customers are unlikely to be affected by the
previous failure of Amazon Restaurants given they may not have been a
member then, and, even if they were, may not have been aware of it as a
result of the [] marketing and awareness of the service.
333
333
In addition to Prime, Amazon’s e-commerce marketplace had almost [] million unique visitors to the UK
store in 2018, with [] million of these ordering at least one item in that year.
167
6.199 In terms of considering the timeframe for Amazon’s re-entry, we are
conscious of the need to avoid spurious claims of accuracy. The CMA’s
guidance does not specify a timeframe that potential entry must occur in for
the purpose of the counterfactual, but that the counterfactual should only
include foreseeable developments.
334
Evidence suggests that five years
would be a reasonable timeframe to expect entry in this specific case. Some
evidence shows entry could be possible within a shorter period of time. First,
Amazon was previously able to enter the UK market within one year of trialling
its offering in the US.
335
Second, third parties indicated that they were able, or
expected to be able (if they decided to do so), to enter the UK market within a
timeframe of a couple of years. Third, given Amazon’s approach of ‘test and
learn’ (see paragraph 6.80), we might expect Amazon to experiment and
launch an offering relatively quickly. This may particularly be the case in a
dynamic market that is continuing to evolve, and where barriers to entry are
expected to have been increasing (see paragraph 6.194). However, the timing
of any entry by Amazon would depend on the route chosen and the
international opportunities that arise, and we have not observed clear
evidence to support an assertion that entry would occur in the very near term.
As such we consider a timeframe of within the next five years to be
appropriate. We consider that Amazon would likely be able to overcome the
technological barriers (such as the development of point-to-point logistics
technology) and commercial barriers (of building customer networks) within
this timeframe, regardless of the route of entry chosen.
Conclusion on acquiring, investing and partnering
6.200 There appear to be multiple possible routes for entry for Amazon and there is
evidence of interest in alternative providers as targets or partners, which could
facilitate Amazon’s entry into the UK market in the short to medium-term (ie
within five years). There exist a number of potential partners and/or targets,
including non-UK restaurant platforms as well as UK-based logistics
specialists, that could help Amazon overcome the barriers to entry to
supplying a restaurant platform in the UK, including those that hampered its
previous attempt in this market. Amazon could use the learning from Amazon
Restaurant to ensure it did not make the same strategic mistakes, or could
invest in or partner to gain additional expertise in this market.
334
MAGs, paragraphs 4.3.6.
335
Evidence relating to Amazon’s attempt in India also indicates it [].
168
Conclusion on re-entry by Amazon counterfactual
6.201 On the basis of the evidence set out above, we believe that Amazon has a
strong and continued interest in online restaurant platforms and an incentive
to offer this to Prime customers in order to differentiate its offering, realise
flywheel benefits and develop useful logistical capabilities that would be
deployed elsewhere in its business. The UK is an important and attractive
market to Amazon as a result of high levels of Prime membership, and
evidence shows that the UK is a large and growing market for online
restaurant platforms. It has demonstrated intention to operate in this market
with its previous entry, Amazon Restaurants and with the decision to invest in
Deliveroo prior to the closure of Amazon Restaurants in the US, which is
consistent with Amazon’s continued intention to be present in this space.
Whilst we have not identified a specific piece of documentary evidence that
unequivocally states Amazon’s clear intention to re-enter the UK market within
a specified timeframe (and we note in that regard Amazon’s general approach
to documentation of its business strategy), we do not consider this
determinative. Indeed, the evidence of incentive and ability, together with the
evidence of Amazon’s long-standing, global interest in online restaurant food
delivery, is sufficient to conclude that the most likely counterfactual is
Amazon’s re-entry.
6.202 We consider the evidence shows there are multiple possible routes for entry
for Amazon absent the Transaction, which include building its own service, or
investing in acquiring, or partnering with an existing online restaurant platform
or an adjacent business. Although much of the evidence relates to potential
target or partner businesses active outside the UK, we consider that the UK is
an attractive market for restaurant delivery (as supported by third party
evidence) and is an important market for Amazon. Whilst entry into the UK
market by Amazon may not occur in the immediate term, particularly because
of Coronavirus (COVID-19) and the uncertainties this is currently causing, we
consider it likely that, absent the investment in Deliveroo, Amazon would look
to re-enter in the short to medium-term (ie within five years).
6.203 We therefore conclude that the most likely counterfactual is one in which
Amazon re-enters the supply of online restaurant platforms in the UK.
7. Supply of online restaurants platforms in the UK
7.1 As described in paragraph 3.16, Deliveroo is active as an online restaurant
platform in the UK as well as various other countries, including Australia,
France, Italy and Spain. Amazon previously operated an online restaurant
169
platform in the UK and the US but it is not currently active in this market
outside of India, where it has recently launched a trial service.
336
7.2 As discussed in the counterfactual chapter, we consider the most likely
counterfactual absent the Transaction is that Amazon would have re-entered
the supply of online restaurant platforms in the UK, which would have led to
the Parties competing with each other in this market in the future.
7.3 In this section, we consider whether Amazon’s investment in Deliveroo could
raise competition concerns in the supply of online restaurant platforms in the
UK as a result of a loss of potential competition.
7.4 We have assessed whether the Transaction would result in an SLC in the
supply of online restaurant platforms in the UK as a result of either:
(a) Amazon using its 16% investment in Deliveroo as its route of entry into
the market, as opposed to entering via another route and competing with
Deliveroo; and
(b) Amazon still re-entering the market in addition to the Transaction, but the
Parties compete less strongly with each other than they would otherwise
compete in the counterfactual. This would be through either: (i) Amazon
competing less strongly with Deliveroo by worsening its own service;
and/or (ii) Amazon influencing Deliveroo to compete less strongly with it
by worsening Deliveroo’s service.
Framework
7.5 The CMA’s Merger Assessment Guidelines provide a framework for the
assessment of unilateral effects arising from the loss of potential competition
as a result of a merger. The guidance sets out two questions to be addressed
when considering actual potential competition:
337
(a) would the potential entrant be likely to enter in the absence of the merger;
and
(b) would such entry lead to greater competition?
7.6 The guidance also states that the CMA will consider whether there are other
potential entrants before reaching a conclusion on the SLC test.
338
336
BBC, 21 May 2020, Amazon trials online food delivery in India.
337
MAGs, paragraph 5.4.15.
338
MAGs, paragraph 5.4.15.
170
7.7 In line with the CMA’s guidance on assessing the loss of potential
competition, we considered (a) whether Amazon is likely to enter, and
(b) would this entry lead to greater competition.
Parties’ submissions
7.8 The Parties submitted that Amazon is not a likely re-entrant in the supply of
online restaurant platforms in the UK. They submitted []. They submitted
[], and that there are high barriers to entry, which have only increased since
Amazon’s previous attempt in the market.
7.9 The Parties submitted that a 16% investment
339
in Deliveroo would not
materially impact Amazon’s incentives on whether (or how) to re-enter
restaurant delivery in the UK. Amazon submitted that its current assessment
is that [] and that, even if this changed, any ‘cannibalisation’ would be
highly unlikely to change any decision to enter (or not). Amazon provided
economic modelling which it submitted illustrates that the influence of a 16%
investment in Deliveroo on Amazon’s decision-making is minimal, and any
incentive or price effects are likely to be less than a tenth of those of a full
merger. Deliveroo further submitted that we are wrong to place any reliance
on the strategic nature of Amazon’s investment, as it is not clear that the
investment is strategic, there can be no expectation by Amazon that it would
have an unfettered path to ultimately acquiring Deliveroo, and any decision by
Amazon to invest in an alternative to Deliveroo would be subject to a cost
benefit analysis.
340
7.10 The Parties submitted that having a 16% investment in Deliveroo would not
cause Amazon to compete materially less aggressively if it did re-enter. This
is because the 16% investment would have only a small impact on Amazon’s
pricing incentives, as only 16% of any diversion to Deliveroo would be
internalised by Amazon.
7.11 The Parties also submitted that there is no basis for the CMA to expect
Amazon, if it did re-enter, to be a particularly close competitor to Deliveroo in
online restaurant platforms.
341
7.12 Finally, the Parties suggested that we need to show it is more likely than not
that each of a series of events and steps would need to occur in order for the
Transaction to result in an SLC, including that it would have acquired,
invested or collaborated with one or more players, that it would enter the UK,
339
The Transaction envisioned Amazon acquiring 16.03% of Deliveroo’s issued share capital and voting rights.
340
Deliveroo response to the Revised Provisional Findings, 10 July 2020, paragraph 4.6.
341
Deliveroo response to the Revised Provisional Findings, 10 July 2020, paragraphs 4.9–4.12.
171
that it would be successful, that this strategy is significantly impacted by the
investment in Deliveroo, and that no other player would enter.
Would Amazon re-enter?
7.13 In order to assess whether Amazon would re-enter the supply of online
restaurant platforms in the UK, we have considered evidence around its
intention, incentives and ability to re-enter. This has been discussed and
described in detail as part of our assessment of the appropriate counterfactual
for our investigation (see paragraphs 6.87 to 6.203). We found:
(a) Amazon has a strong and continued interest in online restaurant platforms
and an incentive to offer this to Prime customers in order to differentiate
its offering, realise flywheel benefits, and develop useful logistical
capabilities that it could deploy elsewhere in its business;
(b) The UK is an important market for Amazon with widespread Prime
membership, and is also a large and growing market for online restaurant
platforms;
(c) Despite closing Amazon Restaurants, evidence shows interest in
investing [] in this area (including the potential [] of Deliveroo).
Moreover, Amazon leadership approved the Transaction [] it decided to
close Amazon Restaurants US, and it has recently launched a trial
restaurant delivery service in India, showing a continued intention to be
present in this sector; and
(d) There are multiple routes of entry providing the ability for Amazon to re-
enter and overcome the barriers to entry. These routes include building
organically, or acquiring, investing, or partnering with an online restaurant
platform business, or one that operates in an adjacent market such as in
last-mile logistics.
7.14 Based on this, we therefore believe that Amazon is a potential entrant who is
likely to enter the supply of online restaurant platforms in the UK in the
absence of the Transaction.
7.15 For the purposes of our competitive assessment, we must compare Amazon’s
re-entry with the situation arising as a result of the Transaction and consider
what impact the Transaction has on the potential future competition between
Amazon and Deliveroo. This assessment must take into account the fact that
the Transaction is the acquisition of a 16% shareholding as opposed to, for
example, the acquisition of a greater shareholding, such as one in which the
acquirer gains full control over the target.
172
7.16 We note that the Parties suggested that we need to show it is more likely than
not that each of a series of events and steps would need to occur in order for
the Transaction to result in an SLC. We consider that this does not accurately
reflect the applicable legal test, which requires the balance of probabilities test
to be applied to the overall question of whether it is more likely than not that
the Transaction leads to an SLC (not to each specific event or step of the
analysis), and that the evidence discussed should be considered in the
round.
342
Theories of harm
7.17 Evidently, if Amazon had acquired 100% of Deliveroo, the acquisition would
result in the full alignment of incentives between Amazon and Deliveroo.
343
On this basis, we would not expect there to be any future competition
between them – ie we may not expect Amazon still to enter in the event of a
full acquisition of Deliveroo, or if we thought Amazon would still enter, we
would expect them not compete with Deliveroo.
7.18 It is possible that a 16% investment could have the same effect as a full
acquisition if the shareholding changes Amazon’s incentives such that it
would no longer enter independently. It is also possible that the investment in
Deliveroo (because the 16% would not change Amazon’s incentives to enter
sufficiently) could lead to a situation where Amazon still were incentivised to
enter independently, leading to Amazon having two competing interests in the
same market.
7.19 Accordingly, in considering whether Amazon would be likely to enter in the
absence of the Transaction, we have considered the impact of the
Transaction on both the binary decision of whether to enter (or not) and the
nature of entry (if a decision to enter is taken). We consider that there are the
two ways in which harm could arise as a result of the Transaction:
344
(a) unilateral effects in relation to Amazon’s decision on entry (meaning that
the 16% shareholding is Amazon’s route to entering, and therefore it
would not enter via another route); and
(b) post-entry unilateral effects, where Amazon still re-enters in addition to
the Transaction, but it worsens one of these offers. This would be either:
342
British Sky Broadcasting Group plc v Competition Commission (CC) (Cases 1095/4/8/08 and 1096/4/8/08)
[2008] CAT 25, paragraph 80.
343
For the avoidance of doubt, acquisition of a smaller stake might also give Amazon the ability to control
Deliveroo.
344
As also identified by the Parties in the CRA’s analysis of implications of Amazon’s 16% holding in Deliveroo
submission.
173
(i) Amazon competing less strongly, to internalise 16% of Deliveroo’s
profits; and/ or (ii) Amazon influencing Deliveroo to compete less strongly
with it.
7.20 We consider that the change in incentives brought about by a 16% investment
is a case-by-case assessment based on the facts and circumstances of a
given transaction and the markets at issue. The outcome of such an
assessment is liable to depend on the nature of the markets in which the
parties to the transaction are active (including the constraints present within
that market and how the market might develop in future), any additional rights
acquired as part of the transaction, as well as the short- and long-term
incentives and intentions of the acquiring firm. We have carried out a
significant amount of evidence-gathering to inform our assessment in this
case. In particular, we have paid close attention to Amazon’s internal
documents to understand the rationale for the Transaction, as well as its likely
longer run expectations and ambitions in the supply of online restaurant
platforms in the UK and internationally.
7.21 In response to our Revised Provisional Findings, Domino’s submitted that we
had wrongly applied a ‘static’ approach to the SLC test by focusing only on
the impact on Amazon’s incentives of the acquisition of a 16% shareholding
as a result of the Transaction. Domino’s submission argues that we should
consider the impact of shareholdings (and additional rights or agreements) up
to the point where Amazon might acquire de facto control of Deliveroo, noting
that the CMA would have jurisdiction to review the acquisition of de facto
control, but not acquisition of additional degrees of material influence.
345
7.22 In support of its arguments, Domino’s referred to statements made by the
OFT in BskyB/ITV that it was appropriate to consider the possibility that
BskyB might be able to appoint a director to the ITV board, or to increase its
shareholding in its substantive assessment of that transaction.
346
7.23 At the outset, we note that we have assessed the Parties’ future plans in
analysing the impact of the Transaction on competition. In particular, we have
carefully considered extensive evidence, including from internal documents
from the Parties and from other Deliveroo shareholders, in assessing the
Parties’ incentives and future plans. We have found that the Transaction is
strategic in nature, not merely financial (see paragraph 3.40) and assessed
345
Domino’s response to the Revised Provisional Findings, 10 July 2020, section 3.
346
Domino’s response to the Revised Provisional Findings, 10 July 2020, paragraph 52.Two other respondents
made similar points, albeit less expansively, see Company D response to the Revised Provisional Findings,
10 July 2020, paragraph 2(b) and Company A response to the Revised Provisional Findings, 10 July 2020,
paragraph 11.
174
how this will affect Amazon’s incentives and future plans, notwithstanding the
fact that the Transaction itself involves the acquisition of a 16% shareholding.
7.24 At the same time, whilst we have found evidence that Amazon might seek to
increase its shareholding in future, we have not observed evidence that would
enable us to form a view of when, or in what form, such an increase might
take place. Nor does Amazon have a [].
7.25 In the absence of such evidence, we disagree with Domino’s submission that
it is incumbent on the CMA to consider the ‘maximum’ version of material
influence that might arise. In the first place, we note that applying the
approach suggested by Domino’s would involve determining all of the multiple
hypothetical scenarios in which de facto control would arise. Yet, as the
CMA’s Jurisdictional Guidance makes clear,
347
determining the existence of
de facto control in future will involve a fact-specific review of whether Amazon
is able to control Deliveroo’s policy in practice, including consideration of any
additional rights and the impact of Amazon’s specific industry expertise as
well as the pattern of voting at board and shareholder meetings. These factors
may, in particular, give rise to de facto control at shareholdings of less than
50%. Any attempt to determine this in advance risks prejudging that
assessment.
348
7.26 At a practical level, given the wide range of possible alternative arrangements,
de facto control might arise at different shareholdings depending on the
additional factors involved. Seeking to define the universe of such
arrangements in this case would, in the absence of additional evidence,
involve speculating not only about Amazon’s future plans, but also about
those of Deliveroo and its other investors (who would have to approve further
investments by or agreements with Amazon).
7.27 Accordingly, we have considered Amazon’s future plans in light of the
acquisition of the 16% shareholding and associated rights through the
Transaction. This approach is consistent with the approach adopted in
reviews of other minority investments.
349
7.28 For completeness, we do not consider that the statements by the OFT in
BskyB/ITV cited by Domino’s require us to take an alternative approach. We
note, in this regard, that the OFT was a phase 1 body and applied a different
legal standard and test to that applied in phase 2. We further note that the
347
CMA2, paragraph 4.28.
348
See, in this regard, section 23(9) of the Act, which provides that the time for determining whether a relevant
merger situation has been created is, in most cases, immediately before the time when the reference has been,
or is to be made.
349
See, for example, RWE AG/E.ON SE (2019), CMA Phase 1 Decision, paragraph 112.
175
target, in that case, was a public company, whose shares were freely traded
and that under the Communications Act 2003 there was a statutory cap in
place on the maximum shareholding BskyB could acquire. This both meant
that an acquisition of additional shares could take place through unilateral
actions by BskyB and that a maximum level of shareholding could be defined.
Impact of the Transaction on whether Amazon would re-enter
7.29 We have first assessed whether the Transaction could result in competition
concerns as a result of Amazon using its investment in Deliveroo as its route
to entry into the UK market, as opposed to entering by another route. This
could lead to a loss of potential competition between Deliveroo and Amazon,
where the latter chooses not to enter the market as a result of the
Transaction.
Parties’ submission
7.30 The Parties submitted that a 16% minority investment would not affect
Amazon’s incentives to enter. In particular, they submitted:
(a) a 16% investment would have a much smaller effect on Amazon’s re-
entry incentives than a ‘full merger’. In particular, while Amazon’s
evaluation of any future entry may be affected by the prospect of this
impacting its investment in Deliveroo, that effect would be substantially
diluted in the context of a 16% investment; and
(b) the effect of a 16% investment is much less likely to stop Amazon from re-
entering a market where it has a strong case for entering (as opposed to a
market where it only has a marginal business case). This is because a
strong entry case would be associated with material profits, such that its
16% investment would be unlikely to change the re-entry decision.
7.31 The Parties submitted that whether or not Amazon would enter is not a binary
question, but rather a question of intensity, with no entry at all as just one
possible outcome.
7.32 The Parties also submitted that it would be ‘illogical to simultaneously contend
both that Amazon would be a strong entrant and that the 16% investment
would make a difference to its decision to re-enter’ ie either Amazon would be
a strong entrant because it had a strong incentive to enter, and therefore the
16% shareholding in Deliveroo would have no bearing on this, or it had a
weak incentive to enter and therefore the 16% shareholding in Deliveroo
would affect this, but Amazon’s re-entry would have had little effect on
competition.
176
Our assessment
7.33 We assessed whether the Transaction could lead to an SLC as a result of
Amazon choosing not to re-enter the supply of online restaurant platforms in
the UK as a result of the Transaction.
7.34 In assessing this, we considered the Parties’ submissions regarding the
changes in entry incentives in the context of a minority shareholding in
comparison with a full merger, including a model the Parties submitted which
sought to quantify this. We then assessed whether, if Amazon viewed the
Transaction as a strategic investment (ie its initial re-entry strategy), its
incentives to enter would change as a result, even in the context of a minority
shareholding. As part of this assessment we considered evidence from the
Parties’ internal documents, views of other shareholders, and Amazon’s
behaviour in other markets.
The Parties’ model of the effects of a 16% investment on the incentives to
enter
7.35 As stated above, the Parties submitted that a 16% investment in Deliveroo
would have a limited effect on its incentive to re-enter the market. As part of
the submission, the Parties shared a theoretical model that they submitted
showed that the overall price effect of a 16% investment is less than a tenth of
the size of a full merger
350
that reduces the number of players from four to
three.
7.36 The model assumed that Just Eat, Deliveroo and Uber Eats were active in all
local markets and that Amazon could choose to enter each market, and if it
enters, to set its price. The model showed that, across a range of
parameterisations,
351
in a market with three incumbent players, the acquisition
of a 16% shareholding in a competitor would have a small impact on the
equilibrium market price
352
(always below 2%), even if it is assumed that a full
merger would have a very significant impact on price (ie up to 20%),
353
and
that therefore in a market with three players, a fourth competitor would not be
350
Full merger is taken to mean Amazon acquiring 100% shareholding/ownership of Deliveroo.
351
Ie a range of choices for the inputs into the model, such as the level of demand, the cost of serving a
customer and the cost of entering a particular location.
352
Ie the market price once all firms had reacted to price changes brought about from unilateral effects of the
merger.
353
The CMA does not have evidence to suggest a full merger would lead to price rises of this magnitude.
177
deterred from entering because it holds a 16% shareholding in a rival unless
the decision to enter was very marginal.
354,355
7.37 We agree that the model is helpful in illustrating the mechanisms at work as a
result of the Transaction and the difference in incentives resulting from a 16%
investment as opposed to a full merger. We note that some of the model’s
specific assumptions, and the calibration of the model, do not accurately
reflect the market (such as the online restaurant platforms being treated as
symmetric, the model not explaining current market outcomes, assuming that
all existing players in the market are active in all local markets and unclear
choice parameterisation).
356,357
As such, we place limited weight on the
results produced by the model, but agree with the logic of the model and
consider it is directionally helpful in illustrating that the effects of making a
minority investment will be different to a full merger.
7.38 We also consider the model provides a static assessment of the financial
implications of the investment in isolation. In particular, it does not account for
the possibility that the investment is strategic in the sense that it gives
Amazon a position from which it can continue to make further investment in
Deliveroo or use its investment to generate a strategic partnership. We
discuss our assessment of this in the following section. However,
notwithstanding the limitations of the model, such that we place limited weight
on the exact results, we do think it illustrates the limited cannibalisation effects
354
An increase in Amazon’s shareholding in Deliveroo would be expected to affect Amazon in two ways. First,
entry into an area by Amazon would divert sales from Deliveroo to Amazon (ie cannibalisation), thus reducing the
return it receives from Deliveroo. If Amazon had a larger shareholding in Deliveroo, this decreasing return would
likely result in Amazon entering fewer areas, Second, if Amazon did enter, the unilateral effects of the merger
would be greater if Amazon had a larger shareholding: a proportion of sales lost by Amazon would be recaptured
by Deliveroo and with a larger shareholding Amazon would likely have an incentive to compete less intensely and
prices rise.
355
In response to the Revised Provisional Findings, Domino’s submitted that entry can take place at different
scale and is not a binary ‘enter or not’ choice, and therefore the 16% shareholding may lead to Amazon entering
in fewer areas (see paragraphs 85 to 90 of Domino’s response to the Revised Provisional Findings, 10 July
2020). However, as described in this paragraph, the Parties’ model considered exactly this point (ie the extent of
the shareholding on the number of areas that Amazon enters and how it competes once it enters – including the
overall effect on price). Given we assessed this as a national market (see paragraph 5.127), and the precise
locations of any Amazon entry would be relatively speculative, we have not assessed Amazon’s entry decisions
in specific local areas. However, we note the general principle holds that if Amazon had a strong incentive to
enter a particular area then it is unlikely a 16% shareholding would materially reduce Amazon’s incentive to enter
that area.
356
The model assumes that firms are all similar along key dimensions, in technical terms symmetric. For
instance, the players in the model are identical with respect to their costs, quality and coverage of areas they are
active. These assumptions may bias the outcomes of the model and we are therefore concerned about the
reliability of the exact numerical results of the model.
357
The model is numerically simulated to illustrate different possible outcomes. We are concerned that, first, this
numerical simulation is based on a model that does not approximate actual market shares. However, responses
of firms may depend on their market share and therefore the model does not take this into account. Second,
evidence suggests that not all players are active in all markets (we observe markets with one, two or three
players). Third, while the Parties claim that they are using a wide and reasonable range in the numerical
simulation for the inputs into the model, we are concerned that those numerical values are not the most realistic
ones. For example, the Parties assume that the highest willingness to pay for an online delivery is a given
amount, however, it is not clear that this reflects the actual maximum willingness to pay by consumers. Overall,
we are therefore concerned about the reliability of the exact numerical results of the model.
178
of a 16% shareholding in comparison with a full merger. As such, if there was
a strong financial incentive for Amazon to enter, it is unlikely that the 16%
shareholding in Deliveroo would materially reduce Amazon’s incentive to re-
enter the supply of online restaurant platforms in the UK.
Amazon’s status as a strategic investor
7.39 We considered whether Amazon may have broader strategic intentions in
making the investment in Deliveroo, which could make it less likely that
Amazon would look to re-enter the market through an alternative route
because it saw the Transaction as part of its re-entry strategy. We have
therefore considered the extent to which Amazon’s investment should be
viewed as a strategic entry route, either as a first step in ownership/control of
Deliveroo, or as a hedge (which allows Amazon to assess the importance of
restaurant delivery in a way it would not be able to do were it not actively
involved with a market participant, like Deliveroo).
7.40 The Parties submitted that Amazon’s investment in Deliveroo was financial
and that it was investing, amongst other reasons, after being impressed by
Deliveroo’s performance and CEO. However, as described in our April
Provisional Findings, we consider the evidence (in particular, Amazon’s
internal documents discussing the Transaction and the restaurant food
delivery industry) demonstrates that Amazon considers Deliveroo to be of
potential strategic value to it in future ([]), and that Amazon’s investment in
Deliveroo was strategic, [].
358
7.41 The evidence we observed showing that the investment in Deliveroo had a
strategic value beyond being merely a financial investment, includes:
(a) Evidence described in paragraphs 6.152 to 6.166 including that it
considered a [] when it invested in Deliveroo, and its rationale for the
[] and the pursued investment in Deliveroo.
(b) In an interview with [] (VP, Corporate Development) he told the CMA
that [].
(c) Emails discussing []. [] (SVP Amazon India, and one of the most
senior executives at Amazon)
359
states that ‘[]’.
360
358
April Provisional Findings, 16 April 2020, paragraph 2.40.
359
GeekWire, 5 December 2019, Amazon expands Bezos’ elite ‘S-team’, adding 6 execs from emerging
branches of the company.
360
[]
179
(d) An internal Amazon email discussing a potential investment in []. In this
example, Amazon’s internal discussion stated it was ‘[]’.
361
7.42 This evidence suggests the investment in Deliveroo could be considered as
Amazon’s route of re-entry as:
(a) Amazon has a strong strategic interest in restaurant delivery;
362
(b) Amazon originally considered [] Deliveroo, noting that it would be hard
to achieve a [] and that Deliveroo ‘[]’, whilst noting the [] of
Deliveroo and broadly agreeing with Deliveroo’s assessment of its [];
(c) The Parties told us that they see the investment as providing the
possibility for further investment in Deliveroo (paragraph 7.46 below);
(d) Amazon’s presence as a shareholder is a deterrent to rivals and future
rivals, and third parties see the investment as Amazon’s re-entry into the
market;
(e) Other Deliveroo shareholders see the Transaction as a possible prelude
to a full acquisition by Amazon, which would be a potential route for those
shareholders to realise their investments in Deliveroo;
363
and
(f) Amazon has a higher liquidation preference, which Amazon describes as
‘[]’; and
(g) Amazon has a right [],
364
which contributes to Amazon’s ability to exert
influence over how a sale of Deliveroo might be achieved.
7.43 However, there are some barriers to Amazon using Deliveroo as its route to
re-entry into the UK market. There would be a cost to Amazon in acquiring or
investing further in Deliveroo (although this may not be prohibitive given
Amazon’s resources), and there are risks that Amazon would not be able to
increase its shareholding in Deliveroo to allow it to use Deliveroo as its route
to re-entry. For example, Amazon risks being outbid by another buyer.
361
Amazon’s acquisition of Ring (which provides doorbell security systems) is another example where Amazon
appeared to invest strategically. It started with a small stake, stating this investment allowed Ring to integrate and
‘lean in’ to Alexa-related capabilities. Amazon then went on to make the full acquisition (see Amazon, 12 April
2018, Amazon and Ring Close Acquisition—Now Working Together to Empower Neighbors with Affordable Ways
to Monitor their Homes and Reduce Crime in Neighborhoods).
362
See April Provisional Findings, 16 April 2020, paragraph 2.32.
363
The evidence of Amazon forming a potential ([]) exit route [] is described in our April Provisional Findings
(paragraphs 3.37–3.44). [].
364
[]
180
7.44 With regard to its interest and incentives in operating an online restaurant
platform, Amazon submitted that [], it would need further integration than a
minority investment. It submitted that these benefits were relevant only [].
365
We consider that the alignment of incentives under a 16% investment is likely
to be less than if Amazon had invested more or acquired Deliveroo, and so
the benefits described above are likely to be harder to achieve. However, we
consider the Transaction gives a degree of alignment, which could give some
ability and incentive for Amazon to pursue strategies that could achieve these
benefits with Deliveroo.
7.45 The Parties’ submission is supported by an Amazon internal email exchange
on [] involving the corporate development team discussing whether
Amazon would be interested in a minority investment in Deliveroo ([]),
which states that:
‘[]’.
7.46 The email described above indicates that Amazon may see the minority
investment as a first step to increasing its shareholding in Deliveroo [].
Amazon told us that it saw the investment as providing ‘a little bit of
optionality’ around investing further or acquiring Deliveroo in the future,
although stated it had no current plans to do so, and that any such further
investment would be subject to CMA review. [] (VP, Corporate
Development) told the CMA that making an investment in Deliveroo was not a
‘[]’. Although the present Transaction involves the acquisition of only a
minority stake, we may consider the extent to which that acquisition also
affects Amazon’s incentives and/or intention to invest in either acquiring or
building an alternative online restaurant platform business, including if it
reflects a decision to use Deliveroo as the route to re-entry.
Whether Amazon would pursue other routes to re-entry
7.47 We also considered whether any future alternative investment by Amazon
could be an independent decision, unaffected by the Transaction. Despite
having invested a significant amount in Deliveroo, Amazon has the resources
and ability to make alternative investments if it had the incentive to do so.
7.48 We agree with the Parties’ submission that if Amazon was presented with a
compelling opportunity to re-enter this market then having a 16% investment
in Deliveroo may not materially affect Amazon’s incentive to undertake such a
365
The rationale stated by Amazon for its minority investment are discussed in paragraph 6.164 and we note that
reference to ‘[]’ and the [] are not included in this rationale, despite being part of the rationale for [] (see
paragraph 6.156).
181
strategy. Its investment in Deliveroo, although significant, results in a minority
shareholding, and so any alternative opportunities in this market would be
subject to a separate cost-benefit analysis, and any cannibalisation effect
could be expected to be small.
7.49 Amazon submitted examples of it entering a sector and cannibalising its own
product line (for example it opened up its e-commerce marketplace to third
party sellers despite this leading to it competing directly with them for sales).
We considered that while these examples show Amazon’s continued
innovation and development of new products and services, they provide
ultimately only weak evidence of Amazon cannibalising its own sales. The
example provided of selling both Amazon Kindle e-readers and books, which
are two substantially different products, is materially different in nature to
offering restaurant food delivery from two online restaurant platform services.
7.50 There is some evidence to suggest that Amazon may have an interest in
pursuing multiple entry routes into the supply of online restaurant platforms:
(a) An email exchange between Jeff Bezos (CEO) and [] (VP, Corporate
Development) in []discussing [] in this space [].
(b) Amazon has continued to pursue the development and launch of a trial for
its own online restaurant platform in India despite investing in Deliveroo.
This involves Amazon investing in building an online restaurant platform,
which it could then look to expand into other jurisdictions, showing
Amazon pursuing both build and buy strategies in the online restaurant
platform market globally.
(c) Amazon conducted discussions with [] during this period, [] (although
Amazon submitted to us that it is not interested in pursuing this
opportunity further).
(d) During email discussions with Deliveroo, [].
(e) Amazon’s modelling of whether to build or buy in the context of deciding
[] Deliveroo implies that building its own service would have a [],
although this modelling was high-level and could be impacted by non-
financial risks as mentioned elsewhere in the analysis (see detailed
commentary in paragraphs 6.158 to 6.162). In the same excel workbook,
Amazon modelled building a restaurant delivery service in the US despite
Deliveroo not being active there (although these cells were hidden). We
note this modelling was not particularly detailed, but still illustrates
Amazon considering the economics of entry and being interested in
launching an online restaurant platform in another jurisdiction.
182
7.51 With reference to its development of restaurant food delivery in India, Amazon
submitted that this shows that Amazon is innovating globally and that ‘it is
fanciful that a 16% minority investment in a single service in a single country
would prevent this deeply held commitment to innovation’ and the potential
application of this innovation to the UK. First, we note that Deliveroo is active
in multiple different countries across the world in addition to the UK, and is
clearly a business offering a service of interest to Amazon. However, we do
agree that despite investing in Deliveroo, Amazon has continued to pursue
the development of an online restaurant platform and, based on its global
approach and examples from other business areas, if this were to result in a
successful restaurant platform, a 16% investment in Deliveroo is unlikely to
deter Amazon from rolling this out internationally. As stated in paragraph
6.102, we see a clear incentive for Amazon to prioritise the UK in any
international rollout.
Conclusion on effects of the Transaction on Amazon’s incentive to re-enter
7.52 We agree with the logic and intuition, as well as the principle of the Parties’
model, that a 16% investment in Deliveroo will have a limited effect on
Amazon’s incentives to enter, when considered statically. However, as a
strategic minority investment, the effect of the Transaction on Amazon’s
incentives to re-enter may be different when viewed dynamically (ie by placing
Amazon in an advantageous position should it wish to increase its holding in
Deliveroo). We consider there is mixed evidence as to whether Amazon’s
strategic investment is likely to materially alter its re-entry incentives. On the
one hand, there is evidence indicating that Amazon views the investment in
Deliveroo as its initial re-entry strategy and could use this investment in the
future to realise its ambition in the UK market, and potentially internationally.
On the other hand, there is evidence indicating that Amazon has an interest in
pursuing multiple routes into the market and that the minority investment in
Deliveroo is unlikely to be sufficient to prevent further investment by Amazon
where a material alternative opportunity arose.
7.53 Based on this evidence, we do not believe it is sufficiently likely that the
investment in Deliveroo would deter re-entry by Amazon if there was a strong
financial incentive for Amazon to re-enter. We note that this assessment could
be different should Amazon acquire a materially larger share in Deliveroo:
should Amazon subsequently look to obtain de facto control, or a controlling
interest in Deliveroo (in order to realise its re-entry strategy), the CMA would
have the opportunity to assess such a transaction.
183
Impact of the Transaction on the nature of Amazon’s entry
7.54 We assessed whether the Transaction could lead to an SLC as a result of
Amazon still choosing to enter (despite having invested in Deliveroo) but
having less incentive to compete strongly against Deliveroo because of the
Transaction, giving rise to horizontal unilateral effects. Horizontal unilateral
effects may arise when it becomes less costly for the merging parties to raise
prices (or reduce quality) as a result of the merger, because they can recoup
the profit from recaptured sales from customers who would have switched to
the other merged party.
366
7.55 As a minority shareholder in Deliveroo, Amazon benefits from 16% of
Deliveroo’s profits. As such, if Amazon loses a sale to Deliveroo, it retains
16% of the profits Deliveroo earns from the sale (but loses 100% of the profits
it would have earned from the sale), while if Amazon wins a sale from
Deliveroo, it captures 100% of the profits from winning the sale but also loses
16% of the profits that Deliveroo would have earned from that sale. This
dynamic could reduce Amazon’s incentive to compete aggressively.
7.56 We considered two scenarios under this theory of harm. In the first, we
assessed whether as a result of the Transaction, Amazon would still re-enter
but would do so in a weaker way (for example not investing as much in re-
entering as it might have done absent the Transaction) leading to a loss of
potential competition that could have been realised if Amazon entered more
strongly. In the second, we assessed whether as a result of the Transaction,
Amazon could influence Deliveroo to worsen its offer, leading to a reduction in
the level of competition that could have been realised if Deliveroo had
competed strongly against Amazon in the future.
• Amazon competing less strongly against Deliveroo
7.57 The Parties submitted that Amazon’s investment in Deliveroo would not cause
Amazon to compete less strongly if it did choose to re-enter because a 16%
investment in Deliveroo would only have a small impact on its pricing
incentives as it would receive only 16% of any benefit Deliveroo received as a
result of it competing less strongly.
7.58 If Amazon were to enter the market and then choose to compete less strongly
against Deliveroo, we would expect Amazon to win fewer customers from
Deliveroo than if it competed strongly. Amazon would also be likely to win
fewer customers from the other incumbents (unless it could avoid competing
366
MAGs, paragraph 5.4.8.
184
strongly against Deliveroo while competing strongly against the other
incumbents). Amazon would miss out on 100% of the profits from any
customer it failed to win from any of the incumbents by not competing
strongly. In contrast, Amazon would gain only 16% of the profits from any of
those customers, retained by Deliveroo, who would have switched to Amazon
if Amazon had competed strongly for them. It is possible that these customers
would be more profitable overall if retained by Deliveroo than if acquired by
Amazon, eg if Deliveroo has lower marginal costs or if the cost to Amazon of
acquiring customers is high. However, we have no evidence to suggest this
would be the case and on balance we would expect Amazon to have a strong
preference for acquiring a customer (and receiving 100% of the profits from
that customer) over allowing Deliveroo to retain that customer (with Amazon
receiving 16% of profits), especially if by competing less strongly Amazon
would also forego winning customers from the other incumbents in which it
does not have an interest.
7.59 We would expect there to be less competitive impact of Amazon entry than if
Deliveroo was the largest player. Just Eat is the current market leader with
more customers (both restaurants and consumers), and so entry by Amazon
is likely to have a greater effect on Just Eat. Therefore, even if competition
was expected to be particularly strong between Amazon and Deliveroo, there
would still likely be material diversion to the other players as a result of
Amazon worsening its offer.
7.60 Even if Amazon was a close competitor to Deliveroo, evidence shows that
Deliveroo, Uber Eats and Just Eat are becoming more similar, both in terms of
their business models (with all three platforms now offering restaurants the
option to use their own in-house delivery or to use the platform’s logistics) and
the restaurants they target.
367
Although restaurant exclusivity may lead to
some differentiation between the online restaurant platforms, evidence
described at paragraph 7.72 shows that restaurants are increasingly moving
towards a multi-homing strategy.
368
7.61 We observed limited evidence that Amazon would be a significantly closer
competitor to Deliveroo than either Uber Eats or Just Eat currently are (thus
leading to higher diversion to Amazon than to either of these competitors).
367
For discussion of business models see paragraphs 5.18 to 5.34 that discuss the move towards hybrid models.
This has allowed the platforms to target the same restaurants. For example, large chain restaurants told us that
they require the online restaurant platform they list on to provide a logistics-enabled service. In addition to this,
we reviewed internal documents provided by Deliveroo, Just Eat and Uber Eats, which showed them competing
to offer a similar restaurant selection (in particular [] which specifically looked to target [] restaurant
partners).
368
Evidence from Deliveroo suggests this has further increased as a result of Coronavirus (COVID-19), where
one internal document noted that approximately [].
185
The closeness of Amazon and Deliveroo would depend on the route to entry
Amazon chooses to take, which would influence what its offering looks like.
Some factors indicating Amazon would be a close competitor to Deliveroo
include:
(a) A Deliveroo survey suggests [].
369,370
This would mean that both would
be targeting and competing to win many of the same customers. In
addition, both Deliveroo and Amazon’s Prime Now
371
offerings
concentrate on London and other major cities, indicating they would also
be competing in the same areas geographically;
(b) Any Amazon offering is likely to include a logistics-enabled platform as
opposed to just a marketplace in which restaurants arrange their own
delivery, based on the importance to Amazon of maintaining control over
the customer experience and the fact Amazon operated logistics as part
of Amazon Restaurants and is doing so in its trial offering in India.
372
Offering logistics to restaurants would make Amazon a closer competitor
to Deliveroo and give it the ability to target the large restaurant chains
who work with Deliveroo and require their restaurant platform to provide
the logistics; and
(c) The objective of Amazon Restaurants was to [], which have been a key
part of Deliveroo’s strategy. [], Just Eat has traditionally targeted []
who have their own delivery capability. This would make the choice of
restaurants available for consumers on the Amazon and Deliveroo
platforms more similar.
7.62 Therefore, we agree with the Parties that it is unlikely the 16% investment in
Deliveroo would cause Amazon to compete materially less aggressively if it
did re-enter because it is not sufficiently likely that Amazon would have the
incentive to worsen its offer in exchange for 16% of the potential benefit that
such a strategy would confer on Deliveroo. We note that this assessment
could be different should Amazon acquire a materially larger share of
Deliveroo.
369
A survey in one of Deliveroo’s internal documents shows that []% of its Plus customers subscribe to
Amazon Prime.
370
But we do not know how this varies for Uber Eats and Just Eat, or with non ‘Plus’ customers.
371
[]
372
In the [] (VP, Corporate Development) phase 1 interview with the CMA, 10 October 2019, he told the CMA
that ‘[]’.
186
Additional concerns raised in responses to the Revised Provisional Findings
7.63 Domino’s submitted that the CMA did not ‘carry out any quantitative analysis
to determine whether Amazon’s incentive to deteriorate its offer is sufficiently
large to result in an SLC’
373
and the ‘CMA’s failure to carry out a GUPPI
374
analysis, given the reliance on GUPPI approaches in many recent merger
cases which are the subject of an in-depth investigation, further undermines
the robustness of its conclusions and its misapplication of the SLC test in the
circumstances’.
375,376
7.64 First, where the CMA has used GUPPIs in cases involving national markets
this is typically as one of several sources of evidence.
7.65 Second, the appropriate evidence base is driven by the facts and
circumstances of each case and the CMA is not required to carry out GUPPI
analysis in order to achieve a robust evidence base. In this case we do not
consider the use of GUPPIs is appropriate. GUPPIs (and other pricing
pressure indices) rely on being able to quantitatively calculate a diversion
ratio.
377
Diversion to Deliveroo will depend on the nature of Amazon’s offer
and its geographic overlap with Deliveroo and, given that Amazon is not
currently in the market (and we do not believe its previous operations would
be reflective of how it would re-enter the market), there is a significant risk that
any estimate (whether from a survey or otherwise) would be inaccurate.
• Amazon discouraging Deliveroo from competing against Amazon
7.66 Finally, we considered whether Amazon’s material influence could enable it to
influence Deliveroo to worsen its offer, resulting in substantially less
competition.
378
Under this scenario, Amazon would recoup 100% of any profit
that arises from sales that are diverted from Deliveroo to Amazon as a result
of this strategy. This means it may have a greater incentive to engage in this
behaviour as opposed to worsening its own offering (where it would only
recoup 16% of any extra profit).
373
Domino’s response to the Revised Provisional Findings, 10 July 2020, paragraph 92.
374
Gross upward pricing pressure index.
375
Domino’s response to the Revised Provisional Findings, 10 July 2020, paragraph 104.
376
As part of its submission, Domino’s pointed to the use of GUPPIs on recent phase 2 cases and calculated
estimates of Amazon’s GUPPI based on different assumptions regarding its gross margin and diversion ration to
Deliveroo (Domino’s response to the Revised Provisional Findings, 10 July 2020, paragraphs 94–100).
377
A diversion ratio is the proportion of all switching customers who would switch from one merging party to the
other.
378
As discussed in detail in the ‘Jurisdiction’ chapter of our April Provisional Findings, the Parties submitted that
the Transaction does not result in Amazon acquiring material influence over Deliveroo.
187
7.67 We considered whether the Transaction could allow Amazon to discourage
Deliveroo from competing against Amazon, for example by causing Deliveroo
to worsen its offer or not pursue strategies that would conflict with Amazon.
For example, it is possible that Amazon’s influence over Deliveroo could
create a perception among Deliveroo’s management and shareholders of the
need to act with caution to avoid conflicting with opportunities that Amazon is
pursuing.
379
Whilst we believe the Transaction will confer on Amazon the
ability to exert material influence over Deliveroo, this influence is less than
would arise in an acquisition of a controlling interest and is not the same as an
ability to control that policy. In particular, it does not amount to an ability to
drive policy in a direction that other shareholders, management or the board
object to. This could make it harder for Amazon to cause Deliveroo to worsen
or reduce its offering if Deliveroo saw this as commercially damaging or
preventing it from engaging in strong growth opportunities.
7.68 In addition, as noted below in paragraph 7.72, we saw evidence of strong
competition between Deliveroo, Uber Eats and Just Eat, which would limit the
scope for Deliveroo to worsen its offer to accommodate Amazon. To the
extent that Amazon were able to influence Deliveroo to worsen its offer,
customers (either restaurants or consumers) lost as a result of any
deterioration of Deliveroo’s offering may not move to Amazon. We would
expect customers to move to Deliveroo’s closest competitors, and so
diversion to Amazon may not be high (depending on whether Amazon’s re-
entry into the market was similar to Deliveroo’s offering).
7.69 Having considered the possible impact of Amazon’s material influence and
the incentives of Deliveroo’s management and shareholders (aside from
Amazon), as well as the specific market conditions, and taking these factors in
the round, we believe that it is not sufficiently likely that the Transaction would
result in Deliveroo competing less strongly against Amazon in the supply of
online restaurant platforms in the UK. We note that this assessment could be
different should Amazon acquire a much larger share of Deliveroo.
Impact of Amazon’s re-entry
7.70 In the course of our investigation, we closely examined the online restaurant
platform market in the UK in order to consider what impact re-entry by
Amazon could have. This included reviewing large volumes of internal
documents from the Parties as well as their competitors, conducting economic
analysis based on data provided by Deliveroo and its competitors, and
379
April Provisional Findings, 16 April 2020, paragraph 3.44.
188
reviewing evidence from restaurant customers and other stakeholders,
including overseas restaurant platforms.
7.71 The Parties submitted that the UK market is currently very competitive and so
any further entry could be expected to have a limited impact. They pointed to
other countries where there are only three platforms active as evidence that
the market may only be able to sustain a small number of players.
7.72 We consider that Amazon would be well-placed to be an effective entrant. It
has the unique asset of its Prime customers, which could allow it to more
easily win customers with lower acquisition costs compared to competitors. It
is well resourced, allowing it to invest in scaling and innovating its offering to
compete effectively. Finally, it has a range of routes available to facilitate entry
(as described from paragraph 6.168).
7.73 Based on the evidence examined during our investigation, we consider that
there appears to be strong competition between Deliveroo, Just Eat and Uber
Eats, but that there is a limited number of suppliers of online restaurant
platforms in the UK. We would generally expect effective entry by a fourth
player to have a positive impact on competition in a market.
7.74 In this market, which continues to grow and evolve, there are some additional
reasons why we believe entry by a fourth player would have a positive effect
on competition:
(a) Although as noted in paragraph 7.60 there is evidence that the online
restaurant platforms are converging (in terms of their business models
and the restaurants they target), the service offered by the online
restaurant platforms is differentiated, with the current suppliers varying in
strength across different parameters of competition eg delivery
capability,
380
commission charged (and other fees),
381
range of
restaurants (for consumers) or consumer base (for restaurants).
382
380
For example, several restaurants highlighted the importance of Deliveroo’s logistics services and told us that
they have not seen Just Eat as suitable for them because it does not provide, or has not previously provided, its
own drivers. Deliveroo’s internal documents noted [].
381
Deliveroo, Just Eat and Uber Eats charge a range of different fees to consumers and restaurants.
Commission rates vary by restaurant and is generally negotiated between the restaurant and the platform. It can
depend on whether the restaurant is part of a large chain, whether it is exclusive to that platform, whether it
requires logistics or whether it operates its own delivery, its size and perceived attractiveness to the platform and
location. Charges to consumers also varies with Just Eat, Deliveroo and Uber Eats charging different fees.
Deliveroo, for example, also offers Deliveroo Plus, a membership scheme where consumers can pay a monthly
fee for unlimited free delivery.
382
Exclusive partnerships, such as the one previously between Uber Eats and McDonalds or the one between
Deliveroo and Wagamama, can drive consumer preference for certain platforms. []. In its documents,
Deliveroo typically focused on large restaurant chains. For example, []. Just Eat told us that it caters for
189
(b) The intensity of competition between the online restaurant platforms
varies between regions in the UK, with Deliveroo being seen as
particularly strong in London.
383
In some areas, there are only one or two
platforms currently operating, and so there may be scope for competition
to increase further.
384
The entry of a fourth national player could therefore
increase competition in some areas from one to two players, or two to
three players.
7.75 A further reason might be the tendency for a market to tip (to a single online
restaurant platform) or tend towards increased concentration, which entry
could counterbalance. The extent of multi-homing we have observed means
that it is not clear to us that the market is likely to tip. Multi-homing by both
restaurants and consumers has become more commonplace in recent years,
with this trend expected to continue.
385
In particular, fewer restaurants are
wanting to form exclusive partnerships with online restaurant platforms and
are increasingly moving towards a multi-homing strategy. Several restaurants
who responded to our questionnaire explained they used multiple online
restaurant platforms because they considered there to be limited consumer
overlap between different platforms.
7.76 Based on our view that we do not believe it is sufficiently likely that the
Transaction will have a significant impact on Amazon’s incentives to re-enter
the market in the UK, or on its incentive to compete against Deliveroo in the
event that it did re-enter, we have not needed to consider these points in
further detail in these findings, or to conclude on the impact re-entry by
Amazon could have in the supply of online restaurant platforms in the UK.
Conclusion on supply of online restaurant platforms in the UK
7.77 Whilst we believe that Amazon is a potential entrant that is likely to re-enter
the supply of online restaurant platforms in the UK, we do not find it
sufficiently likely that the Transaction will have a material impact on Amazon’s
incentives to re-enter, or a material impact on Amazon’s incentives to
compete with Deliveroo in the event of re-entry, such as to result in a
everyone, including those who want something from their local restaurant as well as those who want to order
from a branded restaurant, [].
383
Deliveroo’s strength in London was commented on by third parties and [].
384
Our analysis of geographical coverage based on data submissions from Deliveroo, Just Eat and Uber Eats
highlighted many geographical areas where just one platform (generally Just Eat) operated.
385
For example, in relation to multi-homing by consumers, Deliveroo submitted data []. In relation to multi-
homing by restaurants, Deliveroo submitted that a [] proportion of Deliveroo’s partner restaurants also self-
deliver ([]%), work with Just Eat ([]%), and work with Uber Eats ([]%), as at October 2019 ([]). This
evidence was supported by responses from Just Eat and Uber Eats, who submitted that multi-homing is common
among consumers and certain restaurants. Half of the restaurants who responded to our questionnaire also
confirmed they use multiple online restaurant platforms.
190
substantial reduction in potential competition on the balance of probabilities.
Therefore, we have concluded that the Transaction may not be expected to
result in an SLC in the market for the supply of online restaurant platforms in
the UK.
8. Supply of online convenience grocery services in the
UK
Introduction
8.1 In this chapter we assess the effect of the Transaction on competition in the
supply of online convenience groceries (OCG) in the UK, in which Amazon
and Deliveroo are both present. As described in our Issues Statement,
386
we
assess whether the Transaction will lead to horizontal unilateral effects (ie a
loss of direct competition with a competitor that previously provided a
competitive constraint, or which is expected to provide a competitive
constraint). Horizontal unilateral effects are more likely to arise from a merger
between firms which would otherwise compete closely.
8.2 The OCG market is at an early stage of development; consumers spent
around £[] million on OCG in 2019 (see Table 8.2 below). However, a
number of providers have advanced their current and planned OCG services
in response to the Coronavirus (COVID-19) crises.
8.3 The Parties submitted that Amazon and Deliveroo offer fundamentally
different services, based on different types of logistics networks, and that
closer competition is inconsistent with their respective plans for grocery
delivery.
387
However, we consider that internal documents from each of the
Parties support the view that they may become close competitors in future as
a result of (a) Amazon investing in faster delivery and (b) Deliveroo expanding
and improving its OCG offer.
8.4 The Parties submitted that even if they were to compete in future Deliveroo
would compete closely with its current competitors (the online restaurant food
delivery providers) and Amazon with its competitors ([]) within a highly
competitive sector. We recognise that, if the market grows, other providers will
develop OCG services in competition with the Parties. On the other hand,
both Amazon and Deliveroo have advantages which could enable them to
become market leaders.
386
Issues Statement, 28 January 2020, paragraphs 30–32.
387
Parties’ Initial Submission, 24 December 2019, paragraph 1.6.
191
8.5 We have assessed whether the Transaction would result in an SLC in the
supply of OCG services due to one or more of the following:
(c) Amazon using its influence over Deliveroo to discourage Deliveroo from
competing against Amazon in OCG;
(d) Amazon avoiding competing directly against Deliveroo in OCG, to protect
the value of its investment in Deliveroo; and/or
(e) Amazon relying on Deliveroo to give Amazon a presence in OCG, rather
than developing its own service to compete effectively in OCG.
8.6 The Transaction would give Amazon a 16% holding in Deliveroo. We consider
that the likelihood of the Transaction leading to a SLC through each of the
harms set out above, or a combination of these harms, is less than it would be
in the case of a larger shareholding or a full acquisition. Our view is that the
Transaction will not lead to an SLC in the provision of OCG.
Background
8.7 In this section, we consider: (i) recent trends in the groceries industry, and
(ii) how suppliers compete to supply OCG.
The groceries industry
8.8 UK groceries retailing is a critically important industry which was estimated to
be worth around £190 billion in 2018. Prior to the Coronavirus (COVID-19)
crisis, the industry was expected to grow by around 3% (in nominal terms) per
year to 2023 (see Figure 8.1 below).
8.9 Based on 2018 figures, the majority of grocery sales are generated by large
stores such as hypermarkets and supermarkets (around £106 billion),
388
followed by convenience stores
389
(around £40 billion), discounters
390
(£23 billion), and online groceries (£11 billion) (see Figure 8.1 below).
388
Hypermarkets are defined by IGD as large format stores that sell a full range of grocery items and a
substantial non-food range. Sales areas are typically 60,000 square feet (around 5,500 square metres).
Supermarkets are defined by IGD as food-focused stores with sales areas of between 3,000 square feet (around
280 square metres) and 60,000 square feet (IGD (16 June 2015), UK grocery retailing).
389
Convenience stores are defined by IGD as including store formats typically under 3,000 square feet (around
280 square metres) that sell at least seven core convenience categories. Subsectors include: Symbol groups,
forecourts, convenience multiples, co-operatives and non-affiliated independents (IGD (16 June 2015),
UK grocery retailing).
390
The definition of Discounter which IGD uses includes all sales of Aldi and Lidl as well as grocery-only sales of
bargains stores (Poundland, Poundworld, B&M Bargains, Home Bargains, Wilkinson and Poundstretcher)
(IGD (16 June 2015), UK grocery retailing).
192
Figure 8.1: UK grocery sales, by ‘channel’, 2018 and 2023 Forecast (£billion, %)
Source: IGD (5 June 2018), UK food and grocery market to grow 14.8% by £28.2 billion by 2023.
Notes:
1. ‘Other retailers’ includes specialist food and drink retailers, CTNs (confectionery, tobacco and news), food sales from mainly
non-food retailers and street markets. ‘Discounters’ includes all sales of Aldi and Lidl, and grocery-only sales of principal variety
discounters, including Wilko.
2. The 2023 forecasts were made prior to the Coronavirus (COVID-19) crisis.
8.10 In its recent investigation of Sainsbury’s proposed acquisition of Asda, the
CMA noted several important trends in UK groceries retailing.
391
We consider
the most relevant of these below, along with the potential impact of
Coronavirus (COVID-19) on OCG.
392
Growth of online groceries
8.11 Online sales of groceries were worth £11.6 billion in 2019 according to IGD,
and represented 7.7% of the UK grocery sector. It was forecast to be the
fastest growing grocery channel, expecting to increase in value to £16.7 billion
by 2024, equivalent to a compound annual growth rate of 7.5%.
393
Prior to the
wave of expansion by online restaurant platforms in the supply of online
groceries, and the more recent expansion of OCG service during the
Coronavirus (COVID-19) crisis, Mintel noted that growth of online groceries
overall may be slowing, because ‘online services are still best suited to the
391
Sainsbury’s/Asda, Final Report, 25 April 2019, paragraphs 4.17 and 4.18.
392
In Sainsbury’s/Asda, the CMA also noted the growth of ‘discounters’, ie Aldi and Lidl (Sainsbury’s/Asda, Final
Report, 25 April 2019). Neither is present in the supply of OCG.
393
IGD (20 June 2019), UK food sales to grow by £24 billion by 2024.
105.5 (55%)
112.6 (52%)
40.1 (21%)
47.2 (22%)
23.1 (12%)
31.5 (14%)
11.4 (6%)
17.3 (8%)
10.2
(5%)
9.9 (5%)
2018 2023 (Forecast)
Other
Supermarkets /
Hypermarkets
Convenience
Discounters
Online
Cumulative annual
growth rate, 2018-23F:
-0.6%
8.7%
6.4%
3.3%
1.3%
2.8%Total
£190bn
£219bn
193
traditional big-basket weekly shop, at a time when consumers are increasingly
shopping on a top-up or when-needed basis’.
394
8.12 Despite growth in revenues, suppliers of online groceries have struggled to
make online grocery sales profitable. In March 2019, the Financial Times
reported that the current economics of online delivered groceries are poor, if
not loss making, for the vast majority of grocery retailers.
395
Change in shopping habits
8.13 Prior to the Coronavirus (COVID-19) pandemic, customers of UK groceries
were increasingly shopping ‘little and often’,
396
making around four shopping
trips per week.
397
Within the supermarket (non-discounter) sector, between
2010 and 2018 the percentage of main shop missions declined from 49% to
42%.
398
However, weekly shopping missions remain important for consumers.
Industry reports have found that 89% of customers still do a main weekly shop
where they get all or most of their grocery shopping in one go, although 45%
of customers combine this with other top-up shops.
399
8.14 The Coronavirus (COVID-19) pandemic may have disrupted these trends. In
the initial weeks of the pandemic, supermarkets, convenience stores and
online delivery all saw increased demand.
400
Consumers are also likely to
have changed their shopping habits in the short-term in response to social
distancing (whether favouring convenience stores and online delivery over
supermarket trips, or making fewer and larger supermarket trips to avoid time
spent queuing for entry). However, the medium and longer-term impact of
Coronavirus (COVID-19) on shopping habits remains uncertain.
Growth of convenience
8.15 In line with the gradual trend towards more frequent shopping trips,
convenience store sales (in-store) have been growing in recent years,
representing an estimated 22% of food and grocery sales in the UK in
2019.
401
394
Mintel (12 April 2019), Brits spent £12.3 billion on online groceries in 2018.
395
Financial Times (1 March 2019), The difficulties of making online delivery pay; or KPMG/Ipsos Retail Think
Tank (August 2018), What does the future hold for the UK grocery sector?.
396
IGD (9 February 2016), Three key shopper insights to shape trading with Asda in 2016.
397
Based on Nielsen Homescan data for ‘Average Shops’ for the period ending January 2018 for the UK total
grocers. Provided by Sainsbury’s and Asda during Sainsbury’s/Asda (2019).
398
Based on Kantar data, submitted by Sainsbury’s and Asda in Sainsbury’s/Asda (2019).
399
GlobalData (May 2017), The UK Food & Grocery Market, 2017–2022, page 96.
400
As reported by Kantar (31 March 2020), Record grocery sales as shoppers prepare for lockdown.
401
IGD (20 June 2019), UK food sales to grow by £24 billion by 2024.
194
8.16 In May 2019, Deliveroo forecasted OCG delivery in its markets would []
from £[] in 2018 to £[] in 2021.
402
This is consistent with the views of third
parties expressed prior to the Coronavirus (COVID-19) crisis. For example,
Morrisons told us OCG delivery would grow much more quickly than overall
online groceries sales. Just Eat, Sainsbury’s and Tesco told us that the
consumer desire for OCG would continue to grow, as consumers increasingly
expect to be able to order products at their convenience. Sainsbury’s and
Tesco suggested that the demand for OCG was in part a result of the
expectations consumers have of hot food, being transferred into their
expectations in grocery. However, Tesco told us that it was difficult to forecast
how large the convenience groceries segment will ultimately become.
Effect of Coronavirus (COVID-19)
8.17 The Coronavirus (COVID-19) crisis has had a profound effect on grocery
delivery in the UK, as well as on the grocery sector and the wider economy. In
the short-term, the increase in demand for online grocery delivery has led to
disruption, such as [] for Amazon Prime Now. However most OCG
providers, including the Parties, have expanded their grocery offers in recent
months. Some providers have said that the crisis has accelerated their
expansion of OCG services. We discuss these developments in further detail
below.
8.18 It remains to be seen how Coronavirus (COVID-19) will affect the nascent
OCG market in the medium or longer-term. It is possible that ongoing health
concerns or a change in shopping habits could sustain the uplift in demand for
these services. On the other hand a number of factors could potentially
contribute to a contraction in demand, such as an economic downturn, a
relaxation of social distancing leading to increased use of bricks and mortar
stores, or major grocery providers increasing their capacity in the provision of
online scheduled delivery services.
How OCG services are supplied
8.19 The supply of OCG is a nascent market in which suppliers and potential
suppliers continue to experiment. There is uncertainty about which models will
succeed in this market, and how competition will proceed.
8.20 Supermarkets have been trialling OCG services for several years. An early
example in this space was Tesco Now, which offered delivery of groceries
402
Deliveroo’s forecast reflects its presence across markets outside the UK, although the UK is Deliveroo’s
largest market.
195
within one-hour in central London from a range of 1,000 products, but which
exited the market in 2018. Since 2018, other grocery suppliers have entered
the market, as have suppliers of online restaurant platforms.
8.21 Third parties told us that offering OCG services cost-effectively is difficult.
A persistent challenge is the cost of offering delivery within a few hours, which
is particularly acute given that grocers are struggling to make delivery
profitable even in next day grocery delivery.
403
8.22 Point-to-point and point-to-multipoint delivery networks involve distinct assets
which address these challenges in different ways:
(a) Point-to-point delivery involves taking a single order from a point of origin
(eg a restaurant or grocery store) to a single point of delivery (eg a
consumer) before moving on to deliver another order from a (generally)
different point of origin.
(b) Point-to-multipoint delivery involves taking multiple orders from a point of
origin (eg a logistics hub) and delivering these to multiple points of
destination along a given route. This is sometimes described as the
‘travelling salesman’ or ‘milkman’ model.
404
8.23 We consider below the assets and capabilities used by suppliers of online
convenience groceries, and the extent to which this relates to their use of
point-to-point and/or point-to-multipoint delivery models. Third parties told us
that suppliers of OCG need the following:
(a) a customer-facing e-commerce platform;
(b) locations from which to source groceries;
(c) a courier network; and
(d) dispatch and routing technology.
E-commerce platforms
8.24 Suppliers use customer-facing e-commerce platforms to allow customers to
search, select and pay for baskets of products.
8.25 Grocery retailers told us that an advantage of partnering with online restaurant
platforms is their existing e-commerce platform and customer base, which
403
Mintel (12 April 2019), Brits spent £12.3 billion on online groceries in 2018.
404
Parties’ Initial Submission, 24 December 2019, paragraph 5.29.
196
allows suppliers to market OCG to a targeted audience with a demonstrated
willingness to pay for convenience. However, several grocery retailers told us
that they preferred to use their own e-commerce website, as this gave them
greater control over the customer relationship.
405
One grocery retailer said
this ensured an ongoing element of direct customer contact and control over
the proposition offered (in terms of range, price, promotions and delivery
charges).
Stores and fulfilment centres (‘under the roof’)
8.26 Suppliers of OCG can source groceries from grocery stores or fulfilment
centres. Suppliers can own these stores or fulfilment centres themselves, or
they can partner with third party grocery suppliers. Items can be picked for a
basket either by the delivery courier, or by in-store staff. Grocery suppliers
told us that picking items quickly and efficiently, and reducing the extent of
any substitutions, is a major challenge. For example, [] told us that one of
the main challenges it encountered in operating [] (for which couriers picked
orders) was slow picking and a high rate of substitutions. Co-op told us that it
had temporarily paused its OCG expansion in 2019 to address these
operational issues (although this expansion subsequently recommenced).
8.27 The Parties submitted that small convenience stores are better suited for a
point-to-point network as items are within close reach,
406
while point-to-
multipoint models are suited to ‘central distribution centres/large
supermarkets’, where orders can be batched and picked efficiently. However,
we note that the Parties’ competitors do not necessarily follow one or other of
these models: Ocado Zoom operates point-to-point deliveries (using Stuart)
from its London fulfilment centre, while Sainsbury’s fulfils Chop Chop from its
supermarkets, rather than its convenience stores. In addition, [].
Courier network (‘on the road’)
8.28 Delivering groceries within one- or two-hours of ordering requires a dense
network of couriers who can rapidly fulfil orders. Sainsbury’s and Co-op told
us that scaling a delivery network can be very expensive, while customer
acquisition costs are high.
8.29 The Parties submitted that bicycles and mopeds are a [] component of an
effective on-demand point-to-point model, because they move quickly through
405
Such as []. For example, one grocery retailer told us that a key advantage to self-supply would be greater
ownership of the customer relationship from point of sale to fulfilment.
406
The Parties said this was the case for ‘immediate consumption/impulse shopping missions’. However, the
same logic would apply to other shopping missions that could be met from a convenience store’s range.
197
congested cities, and that this limits the number of items which can be carried
on point-to-point networks relative to point-to-multipoint networks (which use
cars and vans).
8.30 However, we understand that many point-to-point courier networks use a mix
of bicycles, mopeds and cars. Stuart and DoorDash told the CMA that cars
tend to be more suited to rural and suburban geographies, while bikes and
mopeds are best suited to cities.
8.31 Deliveroo’s offer is focused on dense urban areas (primarily []).
407
Prime
Now operates from [] fulfilment centres in nine cities across the UK,
408
and
also fulfils orders from [] Morrisons stores.
8.32 The Parties submitted that []. In contrast, the Parties submitted, point-to-
multipoint deliveries batch orders into single delivery routes [].
409
[] told
the CMA that its point-to-point network could experience scale by batching
two or more orders in the same car once it obtains significant density. We
note that batching is unlikely to be possible using bicycles and mopeds.
Technology
8.33 The distinction between point-to-point and point-to-multipoint networks is set
out in paragraph 8.22 above. The Parties submitted []. [].
410
8.34 Third parties confirmed that these are distinct technologies. Stuart told the
CMA that its point-to-point dispatch technology differs from that of traditional
groceries providers. Co-op also told the CMA that traditional logistics
providers’ technology is not suited to supplying OCG.
8.35 On the other hand, DoorDash told the CMA that its point-to-point technology
allows it to assess and trade off gains in costs from batching multiple
deliveries against the associated reduction in delivery speed and accuracy,
suggesting that logistics models do not necessarily fall neatly into one of the
two categories described by the Parties.
8.36 We note that Amazon has expertise both in operating its point-to-multipoint
operation, and a point-to-point service in the form of [].
407
Deliveroo told us that it is most developed in [] and has had less success in [].
408
In Birmingham, Glasgow, Leeds, Liverpool, London, Manchester, Newcastle, Portsmouth and Sheffield.
409
Parties’ Initial Submission, 24 December 2019, paragraph 5.30.
410
Parties’ Initial Submission, 24 December 2019, paragraph 5.29.
198
Closeness of competition
8.37 Horizontal unilateral effects are more likely to arise when firms compete
closely, as (i) the magnitude of any loss of competition will be greater, and
(ii) each firm is more likely to capture any sales lost by the other as a result of
any worsening in their competitive offer.
8.38 We consider below the likelihood that the Parties will compete in the supply of
OCG in future. In doing so, we consider their:
(a) current OCG businesses and relevant assets;
(b) ambitions in the groceries industry; and
(c) future plans in the supply of OCG.
8.39 We note that while the closeness of competition between the Parties’ existing
offerings is relevant to our assessment, we are principally interested in the
future competition between the Parties.
The Parties’ current OCG businesses
8.40 In this section we compare the Parties’ existing OCG businesses in the UK
and assess their substitutability. Where relevant, we also consider the other
assets and businesses which Deliveroo and Amazon possess which are
relevant for their supply of OCG.
Amazon’s OCG business
8.41 Amazon launched Prime Now in the UK in 2015, before its launch of Amazon
Fresh in 2016. Prime Now operates from [] fulfilment centres across nine
cities, while Amazon Fresh has [] fulfilment centre outside London and
currently makes [] same-day groceries deliveries. In 2016, Amazon began
to supply ‘Morrisons at Amazon’, which offers same-day delivery from
Morrisons using its Prime Now network and Morrisons stores. In response to
the Coronavirus (COVID-19) crisis, Amazon and Morrisons expanded this
service to cover [] Morrisons stores (compared to 16 in 2019). In 2017,
Amazon acquired the US grocery chain Whole Foods, which has seven stores
in London.
8.42 Amazon Prime Now and Fresh had revenues of £[] million in 2019, with
growth of around []% per annum in 2018 and 2019. Groceries accounted
for []% of sales by value, and around []% of deliveries were within []-
hours of ordering.
199
8.43 Amazon told us that Prime Now grocery orders increased by []% between
February and March 2020, but that due to [] its average delivery time []
from around [] hours in February 2020 to over [] hours in March 2020
and [] hours in April 2020. Amazon suggested that Coronavirus (COVID-19)
had, if anything, accelerated the divergence between Prime Now’s grocery
offer and that of Deliveroo. However, Amazon also told us that while
Coronavirus (COVID-19) had led it to make short-term adjustments to its
service, [] of Amazon’s grocery strategy in the UK.
Delivery
8.44 Prime Now is only available to members of Amazon Prime (which costs
£79 per year or £7.99 per month).
411
Prime Now has a minimum order value
of £15, and its delivery pricing varies according to the speed and window of
the delivery, and the size of a customer’s basket (a ‘variable shipping
threshold’ or VST). Amazon Fresh is only available to Prime members in
Greater London, and has a minimum order of £40. Further details of these
offers are shown in Table 8.1. From late July 2020, Amazon began offering
‘free two-hour delivery’ of Amazon Fresh orders above £40 (for Prime
members).
8.45 We note that Amazon withdrew its Prime Now ‘click to deliver’ (CTD) and one-
hour scheduled delivery window options on 16 March 2020. Amazon told us
that it has also temporarily removed Prime Now’s Free Shipping Threshold
(FST) of £40 and started delivering all Prime now orders above the minimum
order value of £15 with no delivery fee. Amazon told us that it made this
change because [].
Table 8.1: []
[]
Source: Amazon.
8.46 Orders are delivered within two-hour windows through the day. Customers
can select a window a minimum of one hour prior to the window starting
time.
412
Accordingly, a customer ordering for the soonest available two-hour
window might guarantee delivery within three hours from ordering, or within
five hours from ordering (if they have just missed the next available
window).
413
411
[]
412
This has recently increased from 45-minutes Scheduled delivery windows are subject to availability.
413
For example, if a customer selects the first window available, an order placed just before 1pm will arrive in the
2pm to 4pm delivery window, ie between one and three hours of ordering. However, an order placed just after
1pm will arrive in the 4pm to 6pm window, ie between three and five hours of ordering.
200
8.47 The Parties submitted that, in 2019, []% of UK Prime Now grocery orders
took more than two hours from order to delivery, just []% took less than one
hour, and just []% were ‘[]’ orders that Amazon committed to deliver [].
The Parties submitted that [] was only available to customers located within
[]-minutes drive of fulfilment centres (whereas two-hour scheduled
deliveries were available to addresses within 45-minutes’ drive) and was
unavailable up to []% of the time for Prime Now and []% of the time for
Morrisons. The Parties submitted in November 2019 that [].
414
It withdrew
this offer on 16 March 2020 in order to improve efficiency in response to the
Coronavirus (COVID-19) crisis.
8.48 Most Prime Now deliveries are made within [] hours of ordering and more
than []% of Prime Now orders are delivered within [] hours from ordering.
This figure rises to around []% including deliveries made by Amazon for
Morrisons at Amazon.
• Amazon Flex
8.49 Amazon Flex is Amazon’s ‘last-mile’ point-to-multipoint logistics network,
made up of independent contractors who deliver Prime Now orders using their
own cars, typically within specified delivery blocks of two to four hours.
415
This
network also fulfils same-day orders for Amazon’s e-commerce marketplace.
As such it delivers both groceries and non-grocery items.
8.50 The Parties submitted that []. However, we note that in the UK Amazon
[]. Amazon Flex also fulfilled Amazon Restaurants deliveries in the UK,
[]. The Parties told us that Amazon Flex [].
• [
]
8.51 [] was launched in October 2017 in the US, as part of Amazon Restaurants
and subsequently rolled out to Amazon Restaurants in the UK. [].
416
[].
8.52 []
414
Parties’ Initial Submission, 24 December 2019, paragraphs 5.7, 5.13 and 5.36.
415
See: Frequently asked questions about Amazon Flex webpage (accessed on 8 June 2020).
416
[]
201
Grocery products
8.53 Prime Now offers a range of approximately [] items from Amazon (including
Whole Foods Market) and a range of approximately [] items from ‘Morrisons
at Amazon’, while Amazon Fresh offers [] items.
8.54 The average Prime Now order contains approximately [] items, while []%
of grocery orders contain at least [] items. Amazon Prime Now’s average
value for orders containing groceries is approximately £[], while its sales
through ‘Morrisons at Amazon’ have an average order value of approximately
£[].
417
The Parties submitted that Prime Now offers groceries at prices [].
Deliveroo’s OCG business
8.55 Deliveroo developed its convenience groceries business as an extension of its
restaurant food delivery business. It has trialled groceries delivery with Co-op,
Nisa and Shell since January 2018, []. However, in late 2018 and early
2019 Deliveroo began to pitch to additional potential partners. These pitches
stated ‘[]’. Deliveroo grew its annual non-restaurant food orders from
around [] in 2018 ([]) to around [] in 2019 ([]).
8.56 Deliveroo told us that it ‘[]’. It said its pitch document was ‘[]’, that
Deliveroo’s grocery team comprises only [] dedicated FTEs with [], and
that its FY2020 plan shows [].
8.57 Deliveroo provides a service through which consumers can access products
from multiple grocery retailers
418
(ranging from retail or symbol group
members to independent convenience stores and off-licences) who can list
convenience items for on-demand delivery. In 2019, Deliveroo’s top five non-
restaurant partners were [].
Delivery
8.58 Groceries purchased on Deliveroo are for immediate delivery, within an
average of around []-minutes and with []% of orders delivered in under
30-minutes.
8.59 Deliveroo’s groceries fee structure is the same as for restaurant food delivery,
including a delivery fee (which varies with the distance over which a delivery
is made), a service fee of £0.49 and a ‘small order fee’ (for orders with value
below a minimum threshold) per delivery. For groceries, Deliveroo charges on
417
Parties’ Initial Submission, 24 December 2019, paragraph 5.11.
418
Although customers cannot order from different retailers in the same order.
202
average a delivery fee of around £[]. Delivery is free (except for the service
fee) for Deliveroo Plus members. [].
Grocery products
8.60 Each of Deliveroo’s partners offers a limited selection of products. (The
following figures relate to ranges offered prior to the Coronavirus (COVID-19)
crisis). For example, Co-op (a key partner which constitutes around []% of
Deliveroo’s OCG sales) offered [] items via Deliveroo. The average range
of a Deliveroo partner was [] items, although some offered over [] items.
Although the total range offered on Deliveroo across its various partners is
large, in any given local area the range of products currently offered on
Deliveroo is limited.
8.61 []
8.62 The Parties submitted that the [] of orders on Deliveroo are [] ([]% of
orders include [] or other age-restricted items),
419
and that of the [] by
quantity in October 2019, [] are [] and [] are []. The Parties further
submitted that [] accounted for around []% of its orders by value (and
[]% by volume) from Co-op in 2019, while other important categories
included [].
8.63 We analysed the list of the 50 items most sold by Deliveroo’s ten largest
partners prior to the Coronavirus (COVID-19) crisis. We find that, consistent
with the Parties’ submissions, [] are the main drivers of demand. On
average, the items in these categories counted respectively for []% of the
total units sold among items listed. Other important categories were []
([]%), [] ([]%) and [] ([]%). Items outside these categories
accounted for only []%.
8.64 The Parties submitted that in 2019, the average Deliveroo order contained
[] items, []% of orders contained one item, and []% of orders contained
five or fewer items.
8.65 Prior to the Coronavirus (COVID-19) crisis, Deliveroo’s average order value
was £[], and []% of its orders were below £15, []% of orders were
below £25 and []% were below £40. Deliveroo noted that its order values
reflected a heavy weighting of [].
419
We note that in the Parties’ Initial Submission, 24 December 2019, paragraph 5.7(ii), the Parties submitted
that the orders including alcohol were []% of the total.
203
8.66 During Coronavirus (COVID-19) Deliveroo’s average order value has risen to
around £[] (April 2020). Deliveroo told us that [] had become more
popular over its platform during Coronavirus (COVID-19). However, []
purchases were still by far the largest category of purchases, with []
accounting for an average of []% of orders by volume across February to
April 2020. In April 2020, [] of the [] items excluding [] were in the []
category, whereas in February 2019, [] of the [] had been [].
8.67 Deliveroo told us that it has expanded its offering in response to an uplift in
consumer demand, arising from the unavailability of grocery delivery slots and
a reluctance among consumers to visit physical stores. Since mid-March
2020, Deliveroo has launched partnerships with BP (M&S food range), M&S
(Standalone), Holland & Barratt, and McColls.
8.68 Morrisons told us that in response to the Coronavirus (COVID-19) lockdown, it
had accelerated its plans with Deliveroo and now services orders from
140 Morrisons stores with a menu of 75 grocery items including fresh food,
groceries and alcohol (compared to one store with a hot food offering at the
start of March 2020). Aldi, told us that prior to Coronavirus (COVID-19) it had
no plans to enter the online convenience grocery space, but that it had started
limited OCG trials with Deliveroo in May 2020.
Substitutability of the Parties’ OCG propositions
8.69 The Parties submitted that Amazon’s grocery delivery service is
fundamentally different from Deliveroo’s service and [].
420
The Parties
submitted that [].
8.70 The Parties submitted that the on-demand convenience ‘top-up’ proposition
offered by Deliveroo is focused on satisfying an emergency or immediate
need and optimized for immediate use/impulse missions.
421
The Parties
submitted that Prime Now is used to [].
422
8.71 We consider below the characteristics of the Parties’ OCG propositions, and
the extent to which these suggest that they are being used by their customers
to fulfil different shopping missions. We consider the Parties’ propositions
from the perspective of both (i) consumers and (ii) grocery suppliers.
420
Parties’ Initial Submission, 24 December 2019, paragraph 5.5. Deliveroo reiterated this point in its response to
our Revised Provisional Findings (Deliveroo response to the Revised Provisional Findings, 10 July 2020).
421
Parties’ Initial Submission, 24 December 2019, paragraph 5.5.
422
Parties’ Initial Submission, 24 December 2019, paragraph 5.5.
204
Competition for consumers
8.72 In this section we compare the main features of the Parties’ OCG
propositions: their delivery service, range and price.
Delivery service
8.73 The Parties have different logistics networks:
(a) Deliveroo’s point-to-point network achieves delivery speeds of less than
30-minutes, reflecting its primary purpose as a restaurant delivery
network. It is limited (compared to Amazon) in the basket sizes and
product ranges it can offer.
(b) Amazon’s point-to-multipoint network largely delivers within several hours
of ordering, with (until recently) a limited amount of one-hour delivery.
Amazon is [].
8.74 Amazon’s Prime Now average delivery speed of around [] hours is much
slower than Deliveroo’s average delivery speed of []-minutes. Prime Now’s
[].
Price
8.75 The Parties submitted (prior to the Coronavirus (COVID-19) crisis) that
Deliveroo’s main groceries partner, the Co-op, prices its groceries on
Deliveroo on average at []% more than Amazon’s prices, based on a
comparison against the top 20 grossing products sold by Co-op on Deliveroo
in 2019 which could be compared. However, our comparison of the top 50
most popular products on Co-op found that Co-op’s prices were [] than
Prime Now, and [] than Morrisons at Amazon.
423
On the other hand, the
average prices of the 50 most popular products sold by Deliveroo’s five
largest groceries partners are [] than those sold on Amazon Prime Now,
and some partners [].
424
8.76 [], as described below. Morrisons told us that [] in its initial trial with
Deliveroo.
8.77 Amazon and Deliveroo use different pricing strategies. Although both Amazon
and Deliveroo charge similar delivery fees, Amazon offers free delivery on
423
The overprice factors have been derived taking into account only the items within the 50 most popular
products sold by Deliveroo’s five largest groceries partners which had an exact correspondence with a product
offered by Amazon.
424
[] and items found on Prime Now accessed on 28 February and on 2 March 2020.
205
orders over £40. Deliveroo customers effectively pay more for building large
baskets (since commission is paid on each item), while Amazon’s customers
are rewarded for building large baskets which exceed the free shipping
threshold. Deliveroo charges [], while Amazon retails its products, typically
at prices []. Amazon also requires customers to subscribe to Amazon Prime
for an annual or monthly fee. We would expect customers to take some
account of subscription and per-delivery fees in evaluating the total cost per
basket delivered (for example when considering whether to re-subscribe to
the service).
8.78 In June 2020, Morrisons told us that in its partnership with Deliveroo (which
has expanded to 140 stores in response to the Coronavirus (COVID-19)
crisis), the prices charged to customers are the same as the instore prices.
[] orders attract a delivery charge of £4.50. [] of the retail value of the
product to Deliveroo as a commission. We note that Deliveroo has [] during
the Coronavirus (COVID-19) crisis ([] from []% prior to 1 March 2020). If
[] in the future, this would likely require [] to change their offer via
Deliveroo.
Range
8.79 We note that Prime Now is also advertised with reference to shorter shopping
missions: ‘Try getting your last-minute items and other household goods’;
‘from a last-minute dinner-party disaster to an emergency diaper situation’; or
‘Twelve guests coming at 6:00 p.m.? And you need a roasting pan because
your leg of lamb is too big for the makeshift pan you usually use? Plus some
wine (in select cities) to get you through dinner prep?’.
425
The Parties
commented that ‘Amazon cannot offer an “ultrafast” on-demand grocery
delivery like Deliveroo and that the language referred to in this paragraph
concerns marketing material, which differs from the documents setting out
details of Amazon’s actual offering’.
8.80 Moreover, Amazon’s internal documents suggest it attempts, among other
things, to meet consumers’ short-term meal needs. For example, [].
8.81 Nonetheless, Amazon’s range is significantly broader than Deliveroo’s. We
therefore note that while some customers’ purchases from Prime Now taking
advantage of its wider range are unlikely to be in competition with Deliveroo,
customers ordering from convenience-focused parts of Amazon’s range may
consider Deliveroo an alternative.
425
See ‘Everything you need to know about Prime Now’ webpage accessed on date 9 June 2020.
206
Competition for grocery suppliers
8.82 In this section, we consider closeness of competition between the Parties in
terms of the route to market which they offer to grocery suppliers.
8.83 The Parties submitted that:
(a) Deliveroo is able to offer retailers genuine on-demand order fulfilment for
a limited range of convenience items and (hot) food deli items (eg in-store
pizza) to be delivered in 20 to 30-minutes from order.
(a) Deliveroo is currently unable to [].
8.84 The Parties submitted that, in contrast to Deliveroo, Amazon can deliver a full
range of goods and offers cold-chain capability, a scheduled service and
warehousing capabilities.
426
The Parties submitted that the service provided
by Deliveroo is largely complementary, rather than substitutable, with the type
of service offered by Prime Now.
427
8.85 We note that Deliveroo currently delivers a range of frozen and chilled
items.
428
Stuart suggested that []. Stuart also told us that it is able to offer
scheduled orders by delaying the point at which an order is entered into its
system.
8.86 [] told us that they can offer scheduled delivery using their point-to-point
delivery networks simply by delaying the point at which the order is fulfilled.
We therefore consider that it would be relatively easy for Deliveroo to add this
capability to its proposition.
8.87 Deliveroo submitted that this would not allow it to offer scheduled delivery in a
scalable way, as delayed orders would require manual action by its operations
team.
429
We consider that Stuart’s example demonstrates these issues could
be addressed using Deliveroo’s existing business model. However, we accept
that Deliveroo’s vehicle fleet limits the size and weight of the grocery basket
which it can currently deliver, and that it therefore cannot fulfil larger orders for
grocers. Addressing this would require Deliveroo to change the composition of
its fleet.
8.88 [] told us that a factor that could make restaurant delivery platforms less
attractive as partners is that this would involve losing control of the customer
426
Parties’ Initial Submission, 24 December 2019, paragraph 5.23.
427
Parties’ Initial Submission, 24 December 2019, paragraph 5.24.
428
The Parties commented that there was a difference between delivering frozen products for immediate
consumption and for storage, but did not explain how this placed different requirements on the supply chain.
429
Deliveroo response to the Revised Provisional Findings, 10 July 2020, paragraph 3.7.
207
relationship. One grocery retailer told us that using an online restaurant
platform would remove them from the direct customer relationship, thereby
limiting their ability to control the quality of the delivery service and conduct
appropriate checks and verifications (eg age checks for alcohol/tobacco
sales). Despite this, the retailer would still bear liability and brand reputational
impacts for any failures in the service. The same factor makes Amazon less
attractive as a potential partner. However, as described below, grocers have
few options for partnering with business-to-business providers to offer OCG.
In practice, grocers have chosen to partner with Deliveroo and Amazon (and
in Morrison’s case, both).
Survey evidence on substitutability
8.89 Deliveroo’s grocery offering is a recent development and to date has been of
limited scale. This makes it difficult to evidence the extent to which customers
of either party see the current offer from the other party as a substitute, and
the extent to which this would change as Deliveroo makes its offer available
more widely.
8.90 The Parties have argued that the two offers serve different shopping missions,
with Deliveroo’s offer serving more urgent impulse purchases and Amazon’s
offer being closer to a weekly ‘big basket’ shop. Impulse items (such as
alcohol and confectionary) account for a large proportion of Deliveroo sales.
8.91 We commissioned an online survey of the Parties’ OCG customers,
430
which
was conducted by Accent and PJM Economics.
431
The survey included
standard questions about how the current services are used and whether they
are seen by customers as alternatives to one another. It also included a
conjoint analysis (discussed below) which was intended to provide evidence
of how future changes to the services offered may affect demand. As regards
the current services, the survey found that:
(a) While each Party had a different ‘typical’ shopping mission, both served a
range of shopping missions. A majority of Amazon customers (55%) had
most recently used the service for general groceries, whereas a majority
of Deliveroo customers (59%) had most recently used the service for
impulse/indulgence goods (eg alcohol, tobacco, confectionary).
However,
11% of Deliveroo customers most recently used the service for general
grocery shopping, and 12% of Amazon customers most recently used the
service for impulse/indulgence goods. Similar proportions of each
430
That is Amazon Prime Now customers who had recently ordered groceries for same-day delivery, and
Deliveroo customers who had recently ordered groceries.
431
Accent and PJM Economics (April 2020), Final Report.
208
customer group had used the service most recently for other shopping
missions such as convenience items (Amazon 17%, Deliveroo 16%), or
items forgotten from a planned shop (Amazon 8%, Deliveroo 9%).
(b) Deliveroo customers typically had a more immediate service need. Three
quarters of Deliveroo customers either said they wanted to consume the
items purchased straight away/as soon as possible (57%) or within an
hour or so (17%), whereas the highest proportion of Amazon customers
said they wanted to consume purchased items the next day or later
(38%). However, around a quarter of Amazon customers wanted to
consume the items either straight away/as soon as possible (20%) or
within an hour or so (6%).
(c) Correspondingly, nine out of ten Deliveroo customers said they needed
delivery within 30-minutes (53%) or 30-minutes to one hour (37%). In
contrast, only one in five Amazon customers said they needed delivery
within either 30-minutes (2%) or 30-minutes to one-hour (17%).
(d) Most Deliveroo customers (59%) reported buying ‘hardly any’ of their total
groceries over the last three months via Deliveroo’s service, compared
with 18% of Amazon customers buying ‘hardly any’ from Prime Now. In
contrast, 42% of Amazon customers reported buying half or more of their
groceries via Prime Now, compared with only 7% of Deliveroo customers
buying half or more from Deliveroo.
(e) Fast delivery was the most commonly mentioned reason for choice of
provider (76% of Amazon customers and 59% of Deliveroo customers
gave it as a reason).
(f) Customers of the two Parties identified a broadly similar set of changes
that would make them use the service for more of their grocery shopping.
These included wider range (Amazon 58%, Deliveroo 47%), free/lower
delivery charges (Amazon 35%, Deliveroo 48%), and lower item prices
(Amazon 32%, Deliveroo 48%). As the figures indicate, delivery charges
and prices were more of an issue for Deliveroo customers but also
important to Amazon customers.
(g) A majority of customers reported that they would not purchase groceries
via OCG if, hypothetically, their provider stopped offering the service, with
the most common response being that they would purchase in-store (31%
of Amazon customers and 37% of Deliveroo customers). Only 11% of
Amazon customers and 20% of Deliveroo customers said they would
have ordered groceries for same-day delivery from another provider.
209
(h) Of the 11% of Amazon customers who said they would have used a
different provider for same-day OCG had Amazon stopped offering the
service, the most frequently mentioned alternatives were Just Eat (13%),
Zoom (13%), Deliveroo (12%), Iceland (10%), Uber Eats (9%) and Chop
Chop (5%), but 28% did not identify an alternative supplier. Of the 20% of
Deliveroo customers who said they would have used a different provider
for same-day OCG delivery had Deliveroo stopped offering the service,
the most frequently mentioned alternatives were Uber Eats (35%), Just
Eat (19%),
432
and Amazon (17%), with other providers each mentioned by
less than 10% of customers and only 8% not identifying an alternative
provider.
(i) The proportion of respondents for whom the other merger party would be
their next best alternative (diversion ratio) was low (the diversion ratio for
Amazon to Deliveroo is 2% and for Deliveroo to Amazon is 5%). However,
this may reflect the relatively limited roll-out and awareness of OCG
services at the time this research was conducted. Indeed the diversion
ratios between Amazon and Deliveroo are comparable to diversion ratios
from Amazon to other providers who provide similar services (Zoom and
Chop Chop).
(j) When asked about their habits over the past 12 months in relation to the
Parties’ wider offerings,
433
just over half (52%) of Deliveroo customers
reported using an Amazon Prime subscription to order non-grocery
goods, while 27% said they had ordered groceries (for same-day or
longer delivery); overall, 69% reported having a subscription with Amazon
Prime. Amongst Amazon customers, a third (34%) had used Deliveroo’s
restaurant/take-away service and a fifth (20%) Deliveroo’s OCG service in
the last 12 months.
8.92 Taken in the round, the survey evidence indicates that Amazon and Deliveroo
tend to be used for different shopping missions, but there is some overlap in
the shopping missions which they serve (and more substantial overlap
amongst the attributes of choice that matter most to customers). The results
indicate that diversion ratios between the Parties, while low, are comparable
to those from each of the Parties to other market participants who have similar
offers, indicating that the diversion rates may reflect limited consumer
432
In fact, Just Eat’s OCG offer is still at a low-level trial stage at only 15 stores. This may suggest that some
respondents confused the online restaurant and OCG offers of Just Eat (and potentially of other online restaurant
platforms).
433
Including, but not limited to, groceries.
210
awareness (or availability) of different OCG offers, in addition to any
differences in the shopping missions they serve.
8.93 The survey was also the basis for a conjoint analysis which assessed the
trade-offs customers may be willing to make between various attributes of
choice
434
and predicted shares of preference for hypothetical future offerings.
The conjoint design and analysis are described in detail in the survey report.
8.94 Among other results, this analysis confirmed that Deliveroo customers placed
higher value on delivery speed than Amazon customers in making trade-offs
between both speed and price level and speed and delivery charge, while
Amazon customers valued range much more than speed.
8.95 The Parties expressed concerns about the validity of our survey, in particular
relating to the comprehensibility of the choices offered for the conjoint
analysis, and the risk that Coronavirus (COVID-19) biased the results. As to
survey comprehensibility, we note that (a) we appointed an independent
expert (Professor Stephane Hess) to peer review the conjoint component of
the survey; (b) Accent conducted cognitive testing of the survey design,
including the choices for the conjoint study, to ensure it was understood by
respondents. We recognise the risk that the survey was influenced by the
Coronavirus (COVID-19) crisis. However we note that Accent considered its
results were unlikely to have been significantly affected by Coronavirus
(COVID-19) because the sample was selected based on customer orders
made in advance of widespread reporting of Coronavirus (COVID-19) cases
in the UK, and the majority of the fieldwork was completed before social
distancing guidelines were introduced (see page 14 of the Accent report).
435
8.96 The Parties also conducted their own survey of their customers. We have a
number of methodological concerns about the conduct of this survey which
mean that the results cannot be compared directly with the CMA survey, in
particular: it was not limited to customers for same-day delivery; only a very
limited number of interviews with Amazon customers were achieved; early
questions referring to alternative options may have biased responses to
diversion questions. In addition, the survey was conducted after social
distancing guidelines were introduced in the UK, which may have biased the
results.
436
434
That is price, delivery charge, range and speed of delivery and the levels of these attributes.
435
Accent and PJM Economics (April 2020), Final Report.
436
Accent and PJM Economics (April 2020), Final Report, page 14.
211
Impact of Coronavirus (COVID-19) on closeness of competition
8.97 We have considered whether the changes Deliveroo has made to its OCG
business during the Coronavirus (COVID-19) crisis may have brought it into
closer competition with Amazon. Deliveroo’s internal document ‘[]’ notes
that:
(a) ‘[]’.
(b) This will be delivered from brands such as Co-op, Morrisons, M&S via BP,
McColls, and independents/specialists.
8.98 In addition, Deliveroo is now offering a ‘Deliveroo Essentials’ service from its
Editions sites.
8.99 These developments could be seen as an extension of Deliveroo’s service
from ‘impulse’ shopping missions into the ‘core grocery’ mission. However, we
have reviewed evidence on the changes Deliveroo have introduced in
response to Coronavirus (COVID-19) and note that the changes do not
include any increase in the ranges (from individual suppliers) or basket sizes
offered (although the range available will vary as Deliveroo develops new
partnerships), and the range offered through Deliveroo Essentials is limited to
around 35 items. We also note that the groceries offered through Deliveroo’s
OCG service prior to the Coronavirus (COVID-19) crisis included non-impulse
essential groceries.
8.100 In addition, Deliveroo’s [] notes at the outset that:
(a) Consumers are primarily confined to their homes, in need of obtaining
basic ingredients to cook meals, as supermarkets struggle to provide a
constant food supply, and restaurants are closed to the public.
(b) Governments have asked Deliveroo to help ensure individuals who are at
home during the Coronavirus (COVID-19) pandemic continue to have
access to food.
8.101 Deliveroo told us that the NHS and Department for Health specifically asked
Deliveroo to increase its selection of grocery providers on the platform given
the lack of delivery slots with existing supermarkets.
8.102 Deliveroo continues to service OCG orders using its restaurant delivery
network. We understand that Deliveroo has not [] in order to provide the
services discussed above, and it is not clear that Deliveroo will have a
commercial incentive to continue providing the same services in the medium-
term, when the circumstances outlined above have changed.
212
Conclusion on current competition between the Parties in OCG
8.103 As described above, the OCG offers of Deliveroo and Amazon are
differentiated in terms of their delivery speed, pricing strategies and range.
Our survey found that Amazon’s customers typically use the service for
general groceries, whereas customers typically use Deliveroo to purchase
impulse/indulgence goods (eg alcohol, tobacco, confectionary). However, our
survey also found overlap between the shopping missions fulfilled by the
Parties’ OCG services, and fast delivery was the most commonly mentioned
reason for choice of provider for customers of both.
8.104 The Parties’ logistics networks use different operational models. However,
Amazon has in the past offered delivery within one hour, and currently does
so in the US using a model more similar to [] (see paragraph 8.121).
8.105 On balance, our view is that the Parties are differentiated competitors which
currently compete only to a limited extent.
The Parties’ ambitions in groceries
8.106 In the following, we consider evidence from the Parties’ internal documents
relating to their plans and ambitions in groceries generally, and then
specifically in OCG. These internal documents predate the Coronavirus
(COVID-19) crisis. However, Amazon told us that adjusting to increased
demand resulting from COVID-19 has not led it to [], and Deliveroo has not
[] as a result of COVID-19, [].
Amazon’s ambitions in groceries
Scale of Amazon’s groceries ambitions
8.107 Internal documents evidence Amazon’s ambitions []. The UK in particular is
seen as a ‘big opportunity’ for Amazon in groceries, []:
(a) [].
437
The Parties submitted that this document set out ‘uninhibited blue-
sky aspirations’ rather than the reality of Amazon’s business, and that it
focused exclusively on the US market, which differs materially from the
UK.
437
[]
213
(b) A March 2019 document from Amazon’s corporate development team
proposed []. When Amazon decided on a minority investment in
Deliveroo, its investment approval document noted that [].
(c) [].
438
Amazon commented that this document was exclusively US-
focused, had become outdated, and in any case represented a ‘stretch-
goal’.
(d) [].
[
] and online delivery plans
8.108 []:
(a) []
(b) []
8.109 []
8.110 []
8.111 Amazon told us that []. We note that the plans included delivery within []
hour, but that they do not describe in detail how this will be achieved, and
[]. [].
Deliveroo’s food ambitions
8.112 Given its rapid growth, Deliveroo engaged [].
8.113 [].
439
[].
8.114 A June 2019 Deliveroo strategy document identified grocery as [].
8.115 Deliveroo told us that it was focused on its restaurant food delivery business
and has not []. It said it had []. The Parties told us that Deliveroo’s
ambition to be the ‘definitive food company’ was []. The Parties submitted
that ‘the CMA has chosen to refer to an early scoping document (from June
2019)’ which did not reflect the subsequent evolution of Deliveroo’s business.
We note that June 2019 is a relatively recent document within the context of
Deliveroo’s groceries business.
438
[]
439
[]
214
The Parties’ plans to supply OCG
8.116 As shown above, both Amazon and Deliveroo internal documents indicate
ambitions to compete in the supply of delivered groceries. We now consider
their plans to compete in the supply of OCG.
Amazon’s OCG plans
8.117 Amazon plans [] online groceries delivery business. Its documents describe
this as a ‘Ultrafast Groceries’ business, []. Amazon told us that its current
plans [] remain unchanged in light of Coronavirus (COVID-19), [].
8.118 []
8.119 []
(a) []
(b) []
(c) []
(d) []
(e) []. The Parties submitted that the CMA inquiry had no impact on
Amazon’s business planning.
(f) []
8.120 In summary, grocery delivery within [] hours of ordering was discussed in a
number of Amazon documents dating up to late 2019.
8.121 []
8.122 []
8.123 Amazon told us that [].
8.124 []. We also note that []
440
(although it has not done so to date).
8.125 The Parties told us that ‘[] is aimed at [] by [] depending on its supply-
side capabilities. It underlines, rather than refutes, the fact that Amazon’s
network of centralized FCs [fulfilment centres] []’.
440
Parties’ Initial Submission, 24 December 2019, paragraphs 5.7, 5.13 and 5.36.
215
8.126 We note that delivery within [] would be substantially faster than current
Prime Now delivery in the UK. It remains to be seen whether such a service, if
made widely available, would compete with sub-30-minute OCG offers.
8.127 []. The same document also notes [].
8.128 [] ([]
441
). The Parties commented that ‘the experimental offerings
referenced are the quickest offerings of Amazon’s supermarket competitors
[]’.
Our view of Amazon’s OCG plans
8.129 Amazon’s Ultrafast Groceries service will supply a wide range of products in
the UK, and for many customer missions we would expect this to compete
with traditional grocers’ online delivered groceries offers. However, we
consider there is evidence that Amazon is likely to compete to offer faster
delivery in OCG as:
(a) Amazon has made repeated and sustained efforts to differentiate itself
from traditional grocers on the basis of delivery speed.
(b) offering delivery within one hour was part of Amazon’s plans as recently
as []; and
(c) Amazon has continued to develop its capabilities to offer rapid delivery in
the US.
Deliveroo’s OCG plans
8.130 As described above, Deliveroo has already entered the supply of convenience
groceries in the UK through several partnerships, including with Co-op, Shell,
BP and local symbol group stores. In a short period, it has already built a
business area generating [] per year in sales, and over £[] in delivery and
other fees, even prior to conducting its strategic planning exercise in 2019.
8.131 As part of its 2019 strategy review, Deliveroo identified an ‘Online
Convenience Grocery Opportunity’ [].
Figure 8.2: Deliveroo overview of grocery competitors
[]
Source: Deliveroo.
441
[]
216
8.132 Amazon’s Prime Now service is [] in the list of UK ‘On Demand’ competitors
(see Figure 8.2), [].
8.133 Deliveroo developed its OCG business case over the summer of 2019. By
August 2019, it planned to address [], as shown in Figure 8.3 below.
Figure 8.3: []
[]
Source: Deliveroo.
8.134 Deliveroo’s planning document stated it would:
(a) []. Deliveroo has made a [] with Co-op, through which Deliveroo
would [] expand its partnership with Co-op []. Co-op has since
confirmed that, although this rollout was paused in 2019, it plans to go
ahead and roll out this business to [].
(b) [].
(c) [].
8.135 In summary, Deliveroo’s internal documents appear to propose a [] of its
online convenience grocery offer, with a broad-based strategy including [],
and a target of achieving revenues of £[].
8.136 Deliveroo told us that these plans did not reflect the views of its executive
team, and that it has []. It said that the potential offers with partners such as
[], Booths, and Holland & Barrett would be focused on food for immediate
consumption and impulse items.
442
It also said that its identification of []
reflected that Deliveroo had not fully understood its own capabilities at that
stage. The Parties submitted that in relation to ‘add-ons’, Deliveroo’s aim was
to [] and as such this was an [].
Competition in OCG
8.137 As we discuss below, a number of players are interested in supplying OCGs
and have been expanding in the market, whether in the form of rolling out
services, conducting trials, or negotiating with potential partners. These
potential market participants include:
(a) Online grocery providers, namely Amazon and Ocado.
442
We note that Holland & Barrett currently offers cooking ingredients, toiletries and other non-impulse items
through Deliveroo.
217
(b) Online Restaurant Delivery marketplaces, namely Deliveroo, Uber Eats
and Just Eat.
(c) Convenience store providers including Co-op, Costcutter, Londis and
Nisa.
(d) Traditional grocery providers such as Sainsbury’s, Waitrose, Tesco and
Asda.
8.138 In considering future competition between Amazon and Deliveroo, the
activities of these other players are relevant in that:
(a) The services available on the market may need to evolve in response to
competition from one another. In particular, providers who currently offer
slower delivery speeds may need to find ways to speed up to compete
with OCGs delivered via point-to-point delivery networks, while in turn
services relying on point-to-point networks may need to expand their
ranges and/or price more competitively. Market participants face
significant challenges in improving their offers in these ways, and also in
rolling out their services. However, several players have told us they are
seeking to improve their offers along these lines. To the extent that the
OCG services offered by Deliveroo and Amazon are subject to these
same competitive forces, they may be compelled to improve their services
in ways that will bring them into closer competition with one another than
at present.
(b) As rival services improve and expand, they may become a competitive
constraint on Amazon and/or Deliveroo. Given the nascent state of the
market, and the impact of Coronavirus (COVID-19), recent market shares
may not be informative of future competition constraints. We consider the
relative strengths of current and potential market players, and on the
nature and scale of their recent expansion/entry plans. In addition, we
note that an OCG service requires a number of elements: an ordering
interface (website and app), a grocery supply chain (to stores and/or
fulfilment centres), and a last mile delivery network. Individual market
participants who do not have all of these elements in a fully-developed
form may be able to compete in the market in partnership with others, or
by sub-contracting elements which they do not have in house.
8.139 Our assessment focuses on OCG services. We have considered the extent of
constraint from grocery and convenience stores (‘bricks and mortar’ stores) on
the OCG market. On the assumption that overall demand and household
budgets for grocery will be broadly stable, growth in demand for OCG is likely
to be largely at the expense of traditional grocers. In order to grow, OCG
218
providers will need to ensure that delivery charges and item prices (relative to
in-store) do not exceed customers’ willingness to pay for fast home delivery.
Strong initial demand for some OCG trials and services indicate that some
operators have been able to develop services that achieve this.
8.140 In the following, we briefly describe the services and plans of current market
participants. These descriptions reflect both the services and plans prior to the
Coronavirus (COVID-19) crisis, and updated information from most providers
about their services and plans in light of Coronavirus (COVID-19).
Online grocery providers
Ocado Zoom
8.141 Ocado has provided an OCG service through Zoom since February 2019.
Zoom offers:
(a) Delivery: within 30-minutes or one-hour, or 30-minute scheduled delivery
windows later the same day;
(b) Range: around 10,000 products;
(c) Prices: matched to Ocado’s wider grocery offer;
(d) Minimum order value: £15; and
(e) Delivery charge: from £1.99 to £2.99.
8.142 Ocado told us that average basket spend on Zoom was between £[] and
£[].
443
Zoom deliveries are fulfilled from a [] warehouse and the service is
available only in West London (Ealing, Acton, Chiswick, Kew and some other
surrounding parts of West London). Ocado estimates Zoom’s reach at []%
of the UK population and recorded revenues of £[] million in 2019 (for
deliveries within four hours of ordering). Deliveries are fulfilled exclusively by
Stuart.
8.143 Ocado told us its first site had performed strongly, with retention and spend
exceeding Ocado’s initial business plan. In February 2020, Ocado forecast
that Zoom would reach []% of the UK population and achieve revenues of
£[] million in 2020. [].
444
443
However, average basket sizes have increased during the Coronavirus (COVID-19) crisis.
444
[].
219
8.144 Ocado has seen demand for Zoom across a wide range of shopping missions:
‘Most orders are ‘small’ or ‘top-up shops’ of £30–£50. We also do
see some ‘full shops’ ([]) and ‘dinner for tonight’ missions, very
few orders we would classify as ‘distress purchases’. OZ over-
trades in impulse/BWS [beer, wine and spirits] and under-trades
in household/core grocery’.
8.145 January 2020 sales data indicate demand for groceries via Zoom across the
full range of grocery categories, although with some differences in the
distribution from the standard Ocado service ([]), as shown in Figure 8.4
below.
Figure 8.4: []
[]
Source: Ocado.
8.146 As it continues to develop Zoom, Ocado said it intended to test different lead
times of On Demand (eg delivered in 90/120-minutes) and sensitivity of
delivery pricing, and to transition to the M&S range, following its partnership
with M&S.
8.147 []:
(a) []
(b) []
8.148 Ocado expects that design improvements will enable future sites to achieve
much greater productivity ([]).
Online restaurant delivery marketplaces
Just Eat
8.149 Just Eat told us that it had only recently launched into on-demand groceries
and is still at a trial stage. It is in early discussions with some UK grocers
including Asda, One Stop (Tesco), Co-op Central England and Nisa and is
currently trialling with Asda and One Stop. It told us that its OCG service had
not grown materially between March and May 2020.
8.150 Orders are fulfilled by Stuart. Just Eat is also developing its own last-mile
logistics network, primarily for restaurant delivery.
220
8.151 Prior to the Coronavirus (COVID-19) crisis, Just Eat was partnered with Asda
to deliver OCG. However, Asda told us (at the start of June 2020) that its
online convenience grocery offer had been stopped due to the demand for
high volumes of products and the limitations brought in with social distancing
as a result of the Coronavirus (COVID-19) crisis. Asda told us that it expects
to reintroduce this offer, and that demand is expected to be high when that
happens.
8.152 Prior to the Coronavirus (COVID-19) crisis, Just Eat’s partnership with Asda
offered:
(a) Delivery: within around 30-minutes, or with scheduled delivery within two
days;
(b) Range: 160 products; and
(c) Delivery charge: £4 (including a 50p service charge).
8.153 Just Eat’s partnership with Asda was fulfilled from five Asda stores in London,
Leeds and Leicester, reaching an estimated []% of the UK population in
2019. Asda forecasted (prior to the Coronavirus (COVID-19) crisis) that [].
8.154 In addition to the services it was trialling with Just Eat, Asda told us (prior to
the Coronavirus (COVID-19) crisis) that [].
8.155 []
8.156 Just Eat is also conducting an initial trial with Tesco One Stop, offering on-
demand (30-minutes to one-hour) deliveries of around 200 products, with
orders mainly fulfilled [].
Uber Eats
8.157 Uber Eats had [] relationships with convenience-focused stores as of
February 2020 including Costcutter ([] stores), Bestway Wholesale (Bargain
Booze, Wine Rack, Select Convenience, Central Convenience Stores, Best
One) ([] stores), McColls ([] stores), and Iceland under the brand ‘Sweet
Justice’ ([] store). In 2019, Uber also agreed to acquire Cornershop, a
grocery-focused operator primarily active in Latin America. Uber Eats told us
that it had seen (approximately) [] convenience/grocery stores on its App
since the start of March 2020.
8.158 Uber Eats partnership with Costcutter offers:
(a) Delivery: within around []-minutes;
221
(b) Delivery charge: variable, dependent on the Costcutter store, customer
location and order value.
8.159 Costcutter told us that its offer through Uber Eats was similar to those of Co-
op and Nisa offered through Deliveroo, in terms of range, basket sizes, item
prices, service price and delivery experience. Costcutter expected to reach
[]% of the UK population through Uber Eats in 2020, with revenues of
£[] million. Costcutter submitted that its ambition is to have over [] on
Uber Eats platform in 2022. The offer reached an estimated []% of the UK
population in 2019, with revenues from the online convenience groceries at
£[] million.
8.160 Costcutter has seen the number of independent retailers signed up to the
Uber Eats App through Costcutter increase from around [] per week in the
period preceding the Coronavirus (COVID-19) crisis to around [] stores per
week as of 1 June 2020. Costcutter told us that it expects that the total online
growth that has been seen during the Coronavirus (COVID-19) crisis will
continue to accelerate but at a slower speed.
8.161 Overall, in 2019, Uber Eats reached an estimated []% of the UK population
with online convenience grocery deliveries, earning revenues of around
£[] million. Uber Eats has also been in discussions about delivering for
additional [] stores.
8.162 As a result of the Coronavirus (COVID-19) crisis, Uber Eats has seen []
growth in the number of grocery/convenience stores listed on its App, with the
number [] between the start of March 2020 and the start of June 2020.
Uber Eats told us that it believes that this growth has been driven primarily by
the increase in consumer demand for online grocery shopping as a result of
Government restrictions and as consumers have sought to avoid in-person
journeys. It told us that it believes that in the long-term, even if some of the
demand that has moved online shifts back to in store as Government
restrictions are eased, there will continue to be a wider offering for customers
to shop online, increasing competition and growth of this segment.
Convenience store providers
Co-op
8.163 In addition to its agreement with Deliveroo, Co-op has been trialling its own
online convenience grocery service since March 2019. Co-op offers:
(a) Delivery: within two hours and scheduled one-hour windows later the
same day;
222
(b) Range: around 2,500 products;
(c) Prices: at the in-store price; and
(d) Delivery charge: £3, or free for orders above £25.
8.164 For almost all of its stores, deliveries are fulfilled by Stuart and Ecargo Bikes
via cars, motorbikes or bikes. A partnership with Starship Technologies
includes seven stores where groceries deliveries are made within two hours in
autonomous delivery vehicles. Co-op has also launched a partnership with
personal shopper service Buymie and will be launching a partnership with []
(another personal shopper service) [].
8.165 Co-op told us that its own on-demand food delivery service is a very similar
proposition to the service it runs in partnership with Deliveroo, with the main
differences being that it offers a wider range of products, and these products
are priced at the same level as in Co-op’s stores. Co-op told us that on the
Deliveroo service, different products tend to sell well compared to its own
service, indicating it is reaching different customers.
8.166 Co-op is planning on extending its own OCG service []. It expects to be
able to achieve one-hour delivery in 2020 and [].
8.167 The number of stores fulfilling Co-op’s OCG service grew from 99 stores to
167 stores between March and April 2020. Co-op told us that this expansion
is in line with its rollout plan from before the Coronavirus (COVID-19) crisis.
8.168 Prior to the Coronavirus (COVID-19) crisis, Co-op forecast that its online
convenience grocery service would expand [].
8.169 In April 2020, Co-op modelled a number of scenarios for how the Coronavirus
(COVID-19) crisis may impact its OCG sales. It forecast that there would be
an additional £[] million in revenues from across its own OCG service and
its sales through Deliveroo in 2020.
Traditional grocery stores
Sainsbury’s Chop Chop
8.170 Sainsbury’s has provided online convenience grocery services through Chop
Chop since September 2016. Currently Chop Chop offers:
(a) Delivery: within one hour and scheduled one hour windows later the
same day;
(b) Range: around 5,000 to 6,000 products;
223
(c) Minimum order value: None; and
(d) Delivery charge: £4.99.
8.171 Prior to the Coronavirus (COVID-19) crisis, Chop Chop had a maximum
basket size of 25 items. However, in order to deal with the increased demand
during the Coronavirus (COVID-19) crisis this was reduced to 20 items.
8.172 As of February 2020, Chop Chop was only available in London, with its
deliveries fulfilled from seven Sainsbury’s retail stores (picked and packed by
store assistants). Each store offered delivery within a three kilometre radius
(covering most of Zones 1 and 2). Sainsbury’s estimated Chop Chop’s reach
at around []% of the UK population with revenues of £[] million in 2019.
8.173 Chop Chop’s footprint has expanded significantly since March 2020, as a
result of increased customer demand for OCG brought about by the
Coronavirus (COVID-19) crisis. In April 2020, Chop Chop added two further
London locations. Since then, it has opened a further four locations in London
and non-London stores across Reading, Manchester, Bristol and Brighton. In
total, Chop Chop has moved from seven London stores in early March 2020
to 19 stores as of 1 June 2020, almost doubling its initial postcode reach.
8.174 In 2019 Sainsbury’s estimated Chop Chop’s reach at around []% of the UK
population with revenues of £[] million. Deliveries are fulfilled by Stuart,
[].
Waitrose Rapid
8.175 Waitrose has offered a convenience grocery service through Rapid since
September 2018. Rapid offers:
(a) Delivery: within two hours or in a same-day one hour window;
(b) Range: around 2,500 products;
(c) Minimum order value: £10; and
(d) Delivery charge: £5.
8.176 Rapid has a maximum basket size of 25 items. Deliveries are fulfilled by
Stuart.
8.177 Prior to the Coronavirus (COVID-19) pandemic, Waitrose deliveries were
fulfilled from ten Waitrose branches – eight within London, one in Hove and
one in Bath. Each store offers on-demand delivery within approximately a
224
three mile radius. In 2019 Waitrose estimated Rapid reached []% of the UK
population, with revenues of £[] million.
8.178 In April 2020, Waitrose rolled out Rapid delivery to [] new shops in
London.
445
Waitrose Rapid order volumes increased from [] in February
2020 to [] in May 2020, with revenues growing from £[] to £[] million
over this period.
Expansion and development of competitive offers
8.179 As customer demand evolves and competitors expand and improve their
offers, Amazon and Deliveroo may come under pressure to respond in ways
that are likely to bring them into closer competition with one another, and on
the other hand competitor expansion may provide an increasing constraint to
Amazon and Deliveroo. In assessing potential competitive dynamics, we
consider the following points:
(a) Perceptions of consumer trends and the potential market;
(b) Market participants’ current perception of their offers and competitors;
(c) Plans and/or intentions to expand in the market;
(d) Plans to develop aspects of the service offered; and
(e) Challenges to expansion and service development.
Perceptions of consumer trends and the potential market
8.180 An Ocado internal strategy document comments that: ‘[w]e launched Zoom to
respond to the market and consumer shifts to immediacy. Zoom is an
immediacy offer launched given the trend to more frequent, smaller baskets
and demand for shorter lead teams [sic]. The launch of an immediacy service
was supported by three clear grocery trends: (i) smaller basket, more
frequent, shops; (ii) spend shifting online; and (iii) demand for shorter-lead
time delivery’.
8.181 Ocado told us that Zoom has performed strongly, with retention metrics and
average spend both exceeding its original business plan.
8.182 Uber Eats told us that the Coronavirus (COVID-19) crisis has resulted in
increased consumer demand, which has accelerated conversations between
retailers and online convenience grocery platforms. Uber Eats told us that this
445
Waitrose told us that [].
225
increase in consumer demand is likely to have some lasting impact on buying
behaviour. However, it expected that in the longer-term when stores reopen,
some consumer demand would move back into stores.
8.183 Co-op expanded its OCG offering after its initial trial showed evidence of
consumer demand. It commented that one longer-term effect of Coronavirus
(COVID-19) could be an increased market size due to changes in consumer
behaviour.
8.184 Sainsbury's submitted (prior to Coronavirus (COVID-19) that it had witnessed
an increase in consumer demand whereby consumers expect to be able to
get what they want exactly when they want it. This includes an aggregation of
missions (order types) from groceries to hot food. The (on-demand)
expectations that consumers have of hot food is now translating into grocery,
and Sainsbury's aims to make sure that it does both activities. Sainsbury's
has seen a generational trend in the way consumers shop and it predicts that
the next generation of shoppers won't shop in the same way current and
previous shoppers have done. This has blurred the lines of competition
between supermarkets and convenience delivery services.
446
On the other
hand, Sainsbury’s told us that it sees the OCG proposition as highly
challenging and expects [].
8.185 Sainsbury’s told us that the Coronavirus (COVID-19) crisis had considerably
accelerated demand for online groceries, including OCG, and supply of OCG
services. It does not believe these levels will return to the pre-crisis position
once the crisis settles.
8.186 Waitrose told us that the Coronavirus (COVID-19) pandemic had increased
consumer demand for OCG. It estimates that the market will accelerate
rapidly in 2020, []. In 2021, it assumes that the overall retail sector will
return to normal and that online convenience will pull back sharply, but to a
level that is above the pre-crisis position.
Beyond this, it assumes that the
OCG market will grow faster than the overall online grocery market, driven by
continued customer behaviour change towards online grocery shopping and
convenience missions.
8.187 Just Eat told us that ‘[g]iven the market for convenience grocery has a value
of £42 billion in the UK and is currently all offline, industry experts are
estimating approx. 10% to move online (source: Bernstein Report, although
this does not specify a timeframe for this shift to online)’.
446
Sainsbury’s sees online restaurant delivery as part of this trend, noting that eating at home used to involve a
takeaway or self-prepared food, but with services such as Deliveroo, consumers are able to eat restaurant food
at home. This competes with Sainsbury's as fewer people are eating Sainsbury's self-prepared food at home.
226
8.188 Morrisons considers that ultrafast delivery will grow in the future and that this
area of the grocery industry will grow faster than other channels.
8.189 Tesco told us that it does not have a view of how large the OCG market will
be. However, it said that the market has grown as customer expectations and
needs have changed.
8.190 Tesco told us that instant delivery is currently a different shopping mission to
next-day delivery. However, in the long-term as the instant delivery market
grows and becomes more of the norm for customers, shopping missions are
likely to converge and customers are likely to shop for small and big baskets
alike through either instant or next day delivery. This is likely to be driven by
the propositions retailers offer in the market.
8.191 Asda told us that the impact of Coronavirus (COVID-19) is likely to accelerate
growth in the online convenience grocery market. It believes the online
convenience grocery market will continue to grow as customers demand
quicker, cheaper and smarter solutions.
8.192 Stuart told us that people have started to become more accustomed to the
idea of ordering through their phones and have expectations of delivery that
are far faster for items other than prepared food. It has in turn become more
common for retailers to view speed as either a potential competitive
advantage
or a necessity.
8.193 Stuart commented that in future, if grocers believe that one-hour delivery is
limited to satisfying urgent needs, then basket sizes will remain small, but the
alternative could be that grocers will identify a new shopping experience for
customers, such as shopping for two or three days, which would require a
larger basket.
8.194 A recent report by Mintel notes that existing online convenience grocery
propositions have been 'small localised or trial schemes, not on the national
level needed to drive a true step change in growth. Indeed even if they were
scaled to be national, if that is indeed even possible, the additional costs
associated with such services would be unlikely to attract many consumers
outside the affluent base who are already most likely to shop online’.
447
8.195 In summary, third parties reported a change in shopping habits to smaller and
more frequent shops, and growing expectations for faster delivery. This is
supported by the fact that some early trials have seen a strong demand
response. However, there is no consensus among third parties as to how
447
Mintel, Online Grocery Retailing UK 2019.
227
significant OCG will become in the context of the broader groceries sector.
This will depend in large part on whether market participants will be able to
sustain offers that will be attractive to consumers.
8.196 Providers who expressed a view generally considered that the Coronavirus
(COVID-19) crisis had accelerated both demand and supply for OCG
services, and would have an impact on consumer behaviour beyond the
short-term, increasing the size of the market.
Market participants’ perceptions of their offers and competitors
8.197 A recent Ocado strategy document indicates that it takes a broad view of
potential competitors in OCG ‘[a] number of competitors from different fields
(traditional supermarkets, delivery specialists, Amazon, start-ups) are
exploring this field, though we don’t believe any are profitable or, Amazon
aside and [sic], have scale in the UK – though Co-Op are expanding rapidly
(Exhibit 1). We are also seeing additional competition from hot food delivery
competitors, and the merger of Just-eats and Takeaway, as well as the
potential merger of Amazon and Deliveroo will drive additional competition’.
8.198 Ocado told us it has seen growth in OCG and competitor activity in response
to Coronavirus (COVID-19). This included:
(a) Sainsbury’s Chop Chop, Co-op quick shop and Waitrose Rapid
significantly increasing capacity and coverage.
(b) Amazon expanding its Prime Now food range in partnership with
Morrisons and press reports that Amazon will be launching ‘Amazon Ultra
Fast Fresh’ which will ‘bring rapid delivery grocery to 40% of UK
households by 2021’.
(c) Online delivery platforms/takeaway players (such as Deliveroo, Just Eat
and Uber Eats) increasing their range of groceries.
8.199 Uber Eats told us it currently offers an effective solution (as shown by the
demand response to its offer) for a somewhat urgent, last-minute,
complementary groceries shopping mission, [].
8.200 In a communications plan for the launch of its partnership with Costcutter,
Uber Eats describes the service as offering ‘[i]ncreasing selection and choice
for users – whether you forgot to pick up everyday essentials like bread and
milk or a simple snack, our partnership with Costcutter will give users access
to everyday staples at the push of a button’. On the other hand, the same
document also notes that ‘[m]ost popular items ordered during the trial
included chocolate, crisps, milk, alcohol and soft drinks’.
228
8.201 An Uber Eats internal document indicates that it views [] as competitors for
its online convenience grocery offer. A ‘convenience store strategy’
presentation refers to a range of competitors including [].
8.202 Co-op told us that when it expanded its relationship with Deliveroo after the
initial trial (which ended in November 2018), it increased the number of
products [], and that one of the aims of the product extension was to
include products for meals, not just supplementary products such as
confectionary or alcohol. In a November 2019 internal document setting out its
own (non-Deliveroo) online customer proposition, Co-op commented that:
‘[]’.
8.203 Co-op told us that its own service offers a wider range of products than
offered via Deliveroo, and these products are priced at the same level as in
the Co-op’s stores. However, delivery times are slower (typically two hours)
than through Deliveroo. With the Co-op’s offering on Deliveroo, prices are a
little bit higher and the product range is smaller.
8.204 Co-op has found that on the Deliveroo service, different products tend to sell
well compared to its own service, indicating it is reaching different customers.
It has found that the ‘missions’ (items to be delivered) tend to vary by delivery
time. For 30-minute deliveries, Co-op has found that the orders often include
items which customers ‘need now’ such as forgotten ingredients. It has found
that customers using its own platform tend to order larger baskets and are
likely to order food for more than one meal or for a couple of days. [], as
Co-op works under the assumption that these services reach different
customers.
8.205 In discussing competitors, Co-op focused on the OCG trials and offers of
major grocery providers []. Co-op told us that among the longer-term effects
of Coronavirus (COVID-19) it had identified an increase in competition, as
main competitors had realised the demand for online convenience groceries
and this will likely form a bigger part of their future strategies.
8.206 Sainsbury’s told us it took a wide view towards who it competes with, and
noted that a common measurement in the industry is ‘share of stomach’.
However, Sainsbury’s also said that on-demand and online delivery represent
two separate infrastructures and support two separate customer needs. 'Next-
hour'/ASAP and 'an hour of your choice' are different customer propositions.
Customers who use an on-demand service need the product within an hour
whereas customers who use online delivery are happy to wait until their
scheduled delivery time.
229
8.207 A grocery retailer told us that the significant increase in customer demand for
OCG services in response to Coronavirus (COVID-19) has generated
considerable competition in the OCG market. It said Amazon and Deliveroo
have taken extensive steps during the crisis to expand their presence in the
OCG market – making it clear that: (i) they are actual, direct and close
competitors in OCG/same day grocery delivery; and (ii) they are each a key
route for retailers to access customers and customers to access groceries in
the OCG space. It was of the view that the crisis has enabled Deliveroo to
entrench a first mover advantage in this space. It said that the vast majority of
activity in the OCG space since 1 March 2020 seemed to have been
undertaken by Deliveroo and Amazon.
8.208 Morrisons saw Amazon as an ultrafast provider.
448
However, Morrisons
expected 30-minute delivery to remain limited to hot food and certain grocery
categories.
8.209 In summary, market participants tended to identify a range of different OCG
providers as current or potential competitors, including the on-demand offers
of online grocery providers (Amazon and Ocado), traditional grocery
providers, and convenience store providers, but also grocery delivery through
online restaurant delivery marketplaces including Deliveroo.
8.210 Some market participants saw OCG served via online food delivery
marketplaces and those served via the grocery provider and/or third-party
logistics providers as serving distinct shopping missions. However, we note
this was a matter of perception, and was not supported by evidence from
commercial interactions or market research. In addition, there is potential for
the distinctions between shopping missions to converge as the market
evolves.
Plans and/or intentions to expand in the market
8.211 Prior to the Coronavirus (COVID-19) crisis, Ocado told us that its plan for
Zoom was to [].
8.212 Due to recent growth in the OCG channel and competitor activity resulting
from the Coronavirus (COVID-19) lockdown, Ocado told us that [].
448
With ultrafast food delivery Morrisons considers that it would be quite difficult for it to do this by itself.
Therefore, partnership with an established player such as Amazon was an attractive way to enter the market.
230
8.213 As of February 2020, Uber Eats had partnered with [] convenience stores
including fascia organisations and independent sites. Around three-quarters of
these arrangements had begun in the past few months.
8.214 Uber Eats told us that it wants to [] OCG, and this was [].
8.215 As a result of the Coronavirus (COVID-19) crisis, Uber Eats has seen []
growth in the number of grocery/convenience stores listed on its App, with the
number [] between the start of March and the start of June 2020.
8.216 Just Eat told us that while it is active in OCG, this was not a strategic priority,
as it is focusing on scaling its logistics platform and attracting brands to its
platform. []. Just Eat is also trialling an offer from Tesco One Stop. One
Stop us told that it expects OCG to grow [].
8.217 In an internal document,
Co-op describes its vision as ‘[t]o be the number one
online convenience retailer in the UK, giving more customers access to our
Co-op and seeks to achieve this through a combination of its own online
platform and fleet, and through expansion of its Deliveroo partnership and
other partners’.
A Co-op planning document states ‘[p]lanning 2020 – we will
continue to expand the offer as demand grows, we’ll also work closely with
Deliveroo as they move to new areas []’. In recent months, Co-op has
increased the number of stores in its Co-op Quick Shop offer from 99 to 167,
in line with its plans prior to the Coronavirus (COVID-19) crisis.
8.218 Sainsbury’s Chop Chop has expanded significantly since March 2020 as a
result of the increased customer demand brought about by the Coronavirus
(COVID-19), and it has plans to operate from 50 stores in 20 cities by mid-
2020. Sainsbury’s told us that this reflects its belief in long-term changes to
customer behaviour and continued demand for OCG brought about by the
Coronavirus (COVID-19). Its forecasts (as of 1 June 2020) show Chop Chop
order volumes increasing from [] in March 2020 to [] in March 2021.
However, Sainsbury’s also said that to ensure reliability of its service [].
8.219 For Morrisons, the main focus was to increase its share of the weekly
shopping online grocery business as it sees a bigger opportunity for growth
there. However, Morrisons has accelerated its plans with Deliveroo and now
services orders from 140 Morrisons stores.
8.220 Waitrose told us that its expansion plans have been accelerated due to
elevated demand experienced under the Coronavirus (COVID-19)
231
pandemic.
449
Waitrose told us that it is also evaluating potential plans to roll
out Rapid Delivery to a further []. However, no concrete decision has been
made at this stage.
8.221 []
8.222 In summary, most current and potential market participants expressed an
interest or intention to enter or expand their presence in the OCG market, and
interest in growing in the market was shared across different types of market
participant – online grocery, online restaurant marketplace, convenience
stores and traditional grocers.
8.223 Third parties varied widely in their ambitions for the OCG market, and in the
nature of their plans and activities. However, as described above, a number of
market participants have accelerated their plans in response to elevated
demand experienced under the Coronavirus (COVID-19) crisis.
Plans to develop aspects of the service offered
8.224 Uber Eats told us that in the longer-term, []. In March 2020, Uber Eats
noted that it was currently seeking to acquire a grocery provider in Latin
America (Cornershop) and that this was subject to regulatory review. [].
8.225 Co-op told us that it aspires to achieve one-hour delivery in the near future,
[]. The challenges to faster delivery include: (a) the time taken by Co-op
works to pick items in store; (b) the load capacity of couriers; and (c) the
technology to help the Co-op predict basket size and plan courier availability.
Co-op told us that it may bring its OCG delivery in-house, either using its
current fleet of home delivery vehicles or leasing new vehicles.
8.226 []. One grocery retailer notes that scale and order density are two essential
components to achieving profitability.
8.227 Stuart said its prediction was that one grocery provider would make a
commitment to offering a high-quality one hour delivery service and that other
market participants would quickly seek to improve their offers in response.
8.228 Asda told us that [].
449
Waitrose told us that it also had to pause the service in Bath, due to operational constraints with the 3PL it
uses for last-mile delivery, taking the total number of stores offering the Waitrose Rapid service to 29.
232
8.229 In summary, market participants are considering various developments of
their activities in OCG, including serving broader grocery needs or increasing
basket sizes ([]), and achieving or launching faster delivery ([]).
Challenges to expansion and service development
8.230 Ocado told us that the model used by some providers of fulfilling orders
through retail outlets would raise problems with stock management as
demand expanded. Ocado planned to expand Zoom by building more
fulfilment centres.
8.231 Co-op told us it had experienced such problems. Its internal documents
confirm this, but also describe ways in which Co-op is seeking to manage the
issue: A September 2019 document states ‘[e]nhanced & integrated reporting
– we’re working with Retail IT to feed Deliveroo MI into our Data Warehouse
so that we can provide comprehensive reporting that will allow us to
understand performance, support stores where needed and provide additional
information to Supply Chain colleagues so that they can also support the
stores and making sure they have all the products they need’. However, an
October 2019 document notes that ‘[t]he decision has been taken to pause
roll-out through November and focus resource on providing coaching and
support to stores in order to help improve rejection rates in stores’.
8.232 Co-op told us that the challenges to developing an OCG service included
developing a platform, last-mile routing, and courier fleet integration. It said
the necessary technology could be bought or used in partnership. For
example, Co-op uses Bringg for last-mile routing, Asda uses Oracle and
Morrisons uses Ocado. The e-commerce platform needs to be integrated with
the courier fleet, which can be challenging if several courier services are used
for different locations or purposes (eg bike vs refrigerated van).
8.233 Fixed costs of self-provision are high, so there are benefits to using a ‘pay per
order’ courier service. Co-op has such an agreement with Stuart. However,
we note that [].
8.234 Sainsbury’s told us that it currently only uses Stuart for the logistics side of its
Chop Chop offering. It had started in-house, but it was not economically viable
to maintain an on-demand delivery fleet at that scale; having dedicated bikes
or vans was far too expensive.
8.235 A groceries retailer identified a number of challenges to achieving a profitable
OCG service, including:
(a) Customer acquisition costs: an expensive marketing campaign was not
feasible.
233
(b) Baskets need to be expensive enough to cover the cost of the courier, but
small enough to be transportable. The groceries retailer currently offers a
maximum [] item basket and this can create issues with the load
capacity of couriers.
(c) Stock management to ensure item availability.
(d) Limited availability of couriers at peak times or in bad weather.
8.236 A groceries retailer also noted the challenge of maintaining an element of
customer contact if using a third-party platform (eg Uber Eats) for delivery, as
opposed to using a logistics provider such as Stuart.
8.237 Tesco told us that the key challenges to its Tesco Now trial were: []. The
proposition was deemed to be unscalable without further technological
development to address these issues.
8.238 []. Stuart said that technology was the most important aspect of its
business, because of the need for speed and accuracy in delivery, and this
included having data for forecasting and modelling demand. In addition,
having the density of orders to attract couriers was also important, and this
makes it an expensive business in the early days.
8.239 Stuart told us that if demand shifted to larger baskets for instant delivery then
Stuart would shift its fleet capacity accordingly, for example by adding more
cars, or by adding trailers to bikes. This could also entail moving away from its
existing model in which all its couriers use their own vehicles.
8.240 Stuart told us it had one huge advantage compared to a marketplace which is
the lack of the customer acquisition and retention costs, and Stuart can then
redeploy that money elsewhere.
8.241 In summary, OCG providers have identified a range of challenges to
achieving fast and reliable on-demand delivery, while managing costs to
enable a profitable and price-competitive offer, including customer acquisition
costs. However, at present most market participants are seeking to expand
and improve their offers, while engaging with possible solutions to these
challenges.
Assessment
8.242 Current and potential market participants largely agree that there is scope for
significant growth in the OCG market, consistent with relatively recent
assessments by both Amazon and Deliveroo as set out above. Market
participants also generally see the Coronavirus (COVID-19) crisis as having
234
accelerated growth in the market. There is no consensus on the likely future
size of this market, which depends on the ability of providers to roll out
compelling offers to consumers, such that consumers will be willing to pay a
premium over instore prices.
8.243 Delivery via online restaurant marketplaces is seen as different from other
shopping missions. This may be because of the speed of delivery, range, or
its relatively high price. We note in this respect that, given the limited scale of
OCG services to date, demand for offers via online restaurant marketplaces
may reflect a lack of alternatives (other than visiting a store), rather than
necessarily implying a strong preference for delivery within 30-minutes over
delivery within one hour.
8.244 Almost all larger players in the categories of online delivered groceries, online
restaurant delivery, convenience stores and traditional groceries are currently
active in trialling or offering OCG services with on-demand delivery of
between 30-minutes and two-hours. A number of potentially important market
participants have increased the pace of their OCG expansion plans in
response to Coronavirus (COVID-19).
8.245 Market participants vary widely in a number of respects, including:
(a) the extent to which they see OCG as a priority relative to their core
business, or to other growth opportunities;
(b) how far they have engaged with the challenges of OCG delivery;
(c) the scale and pace of their current OCG activities; and
(d) their business models.
8.246 However, most market participants share an intention not only to expand their
existing offers but also to address their specific limitations, such as higher
prices, limited range, or relatively slow delivery speeds. In addition, we note
that some services – in particular Ocado, and to a lesser extent Waitrose,
already offer a relatively large range, fast delivery, prices which reflect
scheduled delivery or in-store prices and, in Ocado’s case, a modest delivery
charge. As the market matures, and if these or similar offers become more
widespread, they are likely to put pressure on other market participants to
improve their offers along these metrics.
235
Parties’ incentives to improve offers
8.247 In the counterfactual, both Amazon and Deliveroo are also likely to have
incentives to improve their OCG offers as the market evolves. To begin with
Amazon, we note that:
(a) Amazon has consistently differentiated itself from competitors on its
speed of delivery. Amazon Prime Now had a USP of offering same-day
delivery within several hours while traditional grocers were offering next-
day delivery or same-day delivery for orders placed in the morning and
delivered that evening.
(b) To the extent that growth in OCG occurs at the expense of current
demand for online delivered groceries (rather than traditional grocers or
convenience stores) this would have a particular impact on Amazon’s
grocery services.
(c) An effective nationwide OCG service, offering a wide range, could
potentially expand into other aspects of e-commerce, threatening
Amazon’s core business.
8.248 Amazon has expressed its intention to sell food in a ‘differentiated, compelling
way’, in the context of its acquisition of Whole Foods. Recent internal
documents indicate the strategic importance to Amazon of [] as a
differentiator against its grocery competitors:
(a) []
(b) []
8.249 Amazon’s [] describes a need to [].
‘[]’.
8.250 We note that this comment does not distinguish between the ‘ultra-fast’ offer
from Amazon and those of Ocado (Zoom) and Waitrose (Rapid).
8.251 As regards the threat to Amazon’s e-commerce business, in one email chain
discussing [], an online food delivery company in [], an Amazon
representative states that ‘[]’. The Parties submitted that circumstances in
India were not comparable to those in the UK.
8.252 Amazon has submitted that []. However, we note that a range of other OCG
providers are seeking ways to address similar challenges in order to achieve
delivery speeds of two hours, one hour or faster.
236
8.253 As set out above, we note that Amazon was in fact considering [] in the UK
until very recently (see paragraph 8.119 above). In addition, it is rolling out its
[] offer in the US (paragraphs 8.121 to 8.128). The reason Amazon gives
for rolling out faster delivery in the US but not the UK relates to [], but this
also appears to be an issue in the US (see paragraph 8.127 above).
8.254 In addition, we note that Amazon has begun to trial an online restaurant food
delivery service in India.
450
As we discuss in the context of entry to online
restaurant delivery, partnership or acquisition of a non-UK party with a point-
to-point delivery system is another option available to Amazon in introducing
such a system in the UK. Third parties have told us that the technology
powering their platform and delivery network is generally transferable between
countries.
8.255 Overall, our view is that Amazon has an incentive to compete in the supply of
OCG services and, as such, will have a strong incentive to adapt its business
model in order to ensure its offer is competitive with other market participants.
In addition, as we discuss below, Amazon has a number of advantages as an
OCG provider, which could put it in a strong position if it overcame the
challenges around last-mile delivery.
8.256 Deliveroo has set out ambitious plans for OCG delivery, as described in
paragraphs 8.131 to 8.136, although it has subsequently told us that these
plans do not reflect its current position. [].
Future competitive constraints on Amazon and Deliveroo in OCG
8.257 As the discussion above indicates, a successful roll-out of the services
currently planned by rivals may be expected to impose a degree of
competitive pressure on Amazon, Deliveroo, or both. However, a merger may
still give rise to an SLC where other market participants are competing against
the merging parties. The likelihood that other providers will provide a sufficient
constraint on the merging parties to prevent an SLC will be informed by the
competitive strength of the merging parties relative to the overall market, and
in particular the extent to which the merging parties’ offers are differentiated
from other providers and from one another.
Shares of supply
8.258 Because OCG is a nascent market, and particularly in light of Coronavirus
(COVID-19), historical sales are unlikely to be informative of the relative
450
The Parties submitted that this project does not build materially on the experience from Amazon Restaurants,
as India is a very different market.
237
importance of market participants. Most offers have been introduced relatively
recently, and to date have operated at a very limited geographic scale, but are
supported by participants with expansion plans and a presence in related
markets. In addition, the Coronavirus (COVID-19) crisis has led to a
substantial expansion of some OCG services. The Parties submitted that in
2019, most UK Prime Now grocery orders took more than [].
8.259 Table 8.2 sets out sales data for 2019. [].
Table 8.2: OCG sales (gross merchandise value, £m)
Party
2019
Amazon Prime Now (including Morrisons)*
[]
Co-op
[]
Deliveroo
[]
Ocado
[]
Sainsbury's
[]
Uber Eats**
[]
Waitrose
[]
Total Market
[]
Source: CMA analysis based on data from parties.
Notes:
* [].
** [].
Amazon and Deliveroo’s future market position
8.260 Amazon has a number of advantages which could put it in a strong market
position in OCG, particularly if it were to develop a fast (one or two hour) offer.
In particular:
(a) As a web/app-based service, Amazon already has a significant presence
in online retailing, relative to traditional grocers and convenience stores. It
has a large base (a projected []% of households) of Amazon Prime
customers to whom it could promote its service, reducing the customer
acquisition costs which have been identified as a challenge by some other
market participants. Amazon could also promote its service to non-Prime
Amazon customers.
(b) Prime Now has given Amazon a logistics model which it could use as the
basis for expanding its OCG services. []. The Parties submitted that
Amazon does not currently envisage any [] that would be required to
fulfil faster delivery promises.
(c) Take-up of its OCG service could allow Amazon to attract more
customers to Prime and achieve ‘flywheel’ benefits which would not be
238
available to most other current or prospective providers of OCG.
451
This is
likely to increase the value of customer acquisitions to Amazon relative to
other OCGs who have less scope to achieve such flywheel benefits,
making Amazon a relatively aggressive competitor. The Parties submitted
that additions to Amazon’s business do not always lead to flywheel
benefits, and those that perform badly can have a negative impact on
customers’ perception of Amazon.
(d) In courier acquisition, Amazon has a well-known brand and a model ([])
which could allow it to build a large courier base relatively quickly. In
addition, while its existing couriers mostly use cars, these can also be
used for point-to-point delivery in some circumstances.
8.261 Our assessment of entry by Amazon into the online food delivery market
identified some further points which could give Amazon an advantage in OCG
(whether or not it also entered into online food delivery):
(a) Amazon’s ability to sustain losses while it builds the scale needed to be
successful as well as its ability to invest significantly in marketing and its
customer proposition ([]).
(b) Amazon being well placed to innovate (due to its ‘deep pockets’ and
leadership in technological innovations).
8.262 Amazon also has disadvantages relative to some OCG market participants.
As discussed above, it does not yet have a point-to-point logistics network
that could be used for OCG in the UK, which reduces its ability to cost-
effectively deliver OCG within one or two hours. In addition, while it has a
grocery supply chain this is less extensive than those of some traditional
grocers.
8.263 Deliveroo has some important advantages in OCG supply:
(a) As an app-based service, Deliveroo already has a significant presence in
online retailing, relative to traditional grocers and convenience stores. It
has a large customer base on its app, which is likely to have a similar
demographic to the OCG market. For example, []. The future scale of
this advantage will depend on Deliveroo’s future performance in online
restaurant delivery.
451
The term ‘flywheel’ is used at Amazon to describe something similar to a virtuous cycle, which powers the
business. For example, lower prices lead to more customer visits, more customer visits increase the volume of
sales, and that results in more commission-paying third-party sellers to the site.
[].
239
(b) Deliveroo has described itself as having ‘[]’. This is supported by a
sophisticated and well-developed optimisation technology. A key
challenge to potential competitors in the OCG market is the small number
of suppliers of rapid logistics with a wide geographic coverage in the UK.
The only other substantial point-to-point networks are those of Uber Eats
and Stuart, however a grocery retailer told us that Stuart cannot support
expansion into all areas which it would like to target.
(c) Deliveroo has already begun rolling out its OCG offer to a substantial
number of stores. This could potentially give it an early-mover advantage,
however it remains to be seen whether such an advantage will be
important in this market.
(d) As a delivery-only player in the market, along with Uber Eats and Just Eat
it may be able to rely on its grocery partners to address challenges
around supply chain, stock management, and product picking, while
Deliveroo retains ownership of the customer relationship.
8.264 On the other hand, Deliveroo has some disadvantages in the market
compared to some competitors:
(a) []
(b) []
(c) []
Potential closeness of competition between Amazon and Deliveroo
8.265 The Parties have submitted that their current offers are highly differentiated
from one another. We have considered the scope for customers to see the
services offered by each as substitutes as the market expands, and
particularly as the market evolves and potentially converges on very fast
delivery times, sufficiently wide ranges to allow for most grocery shopping
missions, and prices that are broadly competitive with in-store offers.
8.266 To the extent that OCG offers converge in this way, those of Amazon and
Deliveroo may be in closer competition with one another than with some other
market participants. As noted, both Amazon and Deliveroo are app-based
and/or web-based services, in common with Ocado, Uber Eats and Just Eat,
but unlike traditional grocers and convenience stores, Deliveroo’s online
restaurant delivery offer [], and []. [].
452
This may differentiate
452
A survey in one of Deliveroo’s internal documents shows that []% of its Plus customers subscribe to Prime.
240
Deliveroo to a degree from Just Eat, and from most convenience stores and
traditional grocery stores []. The Parties submitted that the overlap in
customer bases is no evidence that they would compete in future, and cited
survey evidence that Amazon and Deliveroo’s (OCG) customers have
different profiles. We note that this could change if more Prime and Deliveroo
Plus customers become OCG customers, and as OCG offers evolve.
Other competitors
8.267 For a market participant to compete effectively in OCG it requires (a) access
to a customer base; (b) a grocery supply chain (ie delivery of groceries to a
chain of stores and/or fulfilment centres); and (c) a last mile delivery solution.
These service elements may be achieved in partnership with other market
participants, or contracted to a third party.
8.268 Online restaurant delivery providers can most readily access a grocery supply
chain by partnering with an existing convenience store chain or traditional
grocer.
(a) Uber Eats is a close competitor to Deliveroo in the UK, with a similar
business model, a large customer base in online restaurant food, and
operating a courier network of a similar, if slightly smaller, scale. Uber
Eats appears well-placed to compete as the market develops.
(b) Just Eat may also develop into a credible competitor in the market,
building on its large customer base. However, its attractiveness as a
partner to grocery providers may depend in part on its success in building
its own courier network. We have not identified evidence that any other
logistics-enabled marketplaces will be credible competitors.
8.269 Traditional grocers and convenience stores have the advantages of (a) an
existing grocery supply chain, reflecting their presence in traditional and
scheduled online groceries and (b) established customer bases and brands.
Grocers can partner with an online restaurant delivery provider for both
access to a customer base and last-mile delivery. Alternatively, they can sell
to their existing customer bases and employ a third party to provide last-mile
delivery. The majority of grocery retailers offering OCG use Stuart to fulfil their
deliveries, although some are considering or developing in-house solutions.
8.270 As discussed above, the OCG services offered by traditional grocers and
convenience stores are at trial phase or at an early stage of roll-out. Their
ability to develop profitable services at a large scale will be subject to the
general challenges in the market (which also affect the Parties), such as
uncertain demand and willingness to pay for OCG and the resulting need to
241
offer value for money while covering the cost of delivery, and the impact of
Coronavirus (COVID-19). In addition, grocers may face the challenges of:
(a) Limited choice of last-mile delivery solution. Stuart currently offers this as
a ‘white label’ solution allowing the grocer to own the customer interface
and relationship, but alternatives to Stuart appear limited, particularly
outside London while development on an in-house point-to-point delivery
solution is challenging.
(b) Developing a large online customer base. Traditional grocers may be able
to convert existing customers of scheduled delivery services to OCG.
They may also be able to market their OCG services to in-store
customers eg via loyalty card schemes. However, as many of their
existing customers are not regularly using the grocer for online purchases,
grocers may face relatively high customer acquisition costs, and a limited
ability to cross-sell to their existing customers.
Theories of harm
8.271 As set out in the CMA’s Merger Assessment Guidelines,
453
in formulating
theories of harm, the CMA will consider how rivalry might be affected by the
merger. For some mergers, the CMA may consider several theories,
sometimes affecting the same market. The CMA will determine whether an
SLC arises or is expected to arise from a merger, having considered one or
more theories of harm. It need not determine this in respect of each of the
theories considered and its overall expectation of an SLC may be based upon
one theory only or upon its composite view of multiple alternative theories.
8.272 As noted in our Issues Statement, we have considered whether the
Transaction may give rise to horizontal unilateral effects in the supply of OCG
services. We have focused on three theories of harm:
(a) Amazon using its influence over Deliveroo to discourage Deliveroo from
competing against Amazon in OCG.
(b) Amazon avoiding competing directly against Deliveroo in OCG, to protect
the value of its investment in Deliveroo.
(c) Amazon relying on Deliveroo to give Amazon a presence in OCG, rather
than developing its own service to compete effectively in OCG.
453
MAGs, paragraphs 4.2.3–4.2.6.
242
8.273 Several third parties have raised additional theories of harm relating to OCG
services and these are discussed in Chapter 7.
8.274 Before discussing each of the above theories of harm in turn, we comment on
two issues which are of general relevance to the assessment: (a) potential
Amazon re-entry in restaurants and (b) future competition in the OCG market,
particularly in the light of the Coronavirus (COVID-19) crisis.
Potential Amazon re-entry in restaurants
8.275 We have set out our views on future Amazon re-entry into the online
restaurant platforms market:
(a) We conclude that the most likely counterfactual for our assessment of
online restaurant platforms is one in which Amazon re-enters the supply
of online restaurant platforms in the UK (see paragraph 6.203).
(b) If there were a strong financial incentive for Amazon to enter, it is unlikely
that the 16% shareholding in Deliveroo would materially reduce Amazon’s
incentive to re-enter the supply of online restaurant platforms in the UK
(see paragraph 7.38).
8.276 Entering the online restaurant platforms market would require Amazon to
have access to a logistics network which could compete with existing online
restaurant platforms (eg by offering delivery in under 30-minutes). As
discussed in paragraph 8.313 below, an Amazon strategy document []
considered the possibility that Amazon would ‘[]’. We note that all three
incumbent online restaurant platforms currently offer OCG services using the
same logistics solutions as for their restaurant delivery services. We therefore
expect that if Amazon entered as an online restaurant platform it could also
potentially provide some or all of its OCG services over the same platform and
logistics network as for restaurant delivery.
454
Future competition
8.277 In order to assess whether any of our theories of harm is likely to lead to an
SLC, we have assessed (amongst other things) the likelihood, absent the
Transaction, of Amazon and Deliveroo emerging as strong players in the
provision of OCG services, and acting as important competitive constraints on
one another. We have also assessed whether the development of the other
454
It remains to be seen how Amazon might configure a combined restaurants and OCG solution in this scenario.
For example, Amazon could integrate such a service with its existing logistics network or operate both
independently, and it could supply groceries through its existing grocery supply chain or through partnerships
with other grocery providers, or both.
243
market participants may be timely, likely and sufficient to prevent any SLC
from arising.
8.278 As discussed in paragraphs 8.244 to 8.246, we note that:
(a) Almost all larger players in the categories of online delivered groceries,
online restaurant delivery, convenience stores and traditional groceries
are currently active in trialling or offering OCG services with on-demand
delivery of between 30-minutes and two-hours.
(b) Most market participants share an intention to expand their existing offers
and address their specific limitations.
(c) Some services already offer a relatively large range, fast delivery, and
prices which are competitive with in-store offers.
8.279 In addition, a number of OCG services are currently in a process of relatively
rapid expansion, in some cases prompted by the Coronavirus (COVID-19)
crisis. In particular:
(a) [];
(b) Uber Eats has seen [] growth in the number of grocery/convenience
stores listed on its App, with the number [] between the start of March
2020 and the start of June 2020 (paragraph 8.215);
(c) Tesco One Stop anticipates [] (paragraph 8.216);
(d) Co-op is continuing the rapid expansion of its Quick Shop offer
(paragraph 8.217);
(e) Sainsbury’s Chop Chop has expanded its footprint significantly since
March 2020, as a result of the increased customer demand brought about
by the Coronavirus (COVID-19) crisis. It plans to operate from 50 stores in
20 cities by mid-2020 (paragraph 8.218);
455
and
(f) Waitrose has expanded its OCG offer in response to growth in demand
during the Coronavirus (COVID-19) crisis and is evaluating options for
further expansion (paragraph 8.220).
8.280 While the future development of the market remains uncertain, we consider
that on the basis of current evidence other market participants appear well-
455
Although it notes that doing so [].
244
placed to compete in OCG provision, particularly in light of the expansion
plans and activities set out above.
8.281 As regards the future market position of Amazon and Deliveroo, as described
in paragraphs 8.260 to 8.264, each of the Parties has advantages which could
mean they are well-placed to compete in OCG services, although each also
has disadvantages relative to certain other market participants.
8.282 A groceries retailer submitted that Amazon and Deliveroo had each taken
extensive steps to expand their position in the OCG market during the
Coronavirus (COVID-19) crises, that the crisis had enabled Deliveroo to
entrench a first mover advantage in this space, and that post-Transaction this
would place the combined entity in a considerable position of power and
weaken the incentive of both Amazon and Deliveroo – who were two key
drivers of competition in this space – to innovate and invest to improve their
respective service offerings.
8.283 Changes to Amazon and Deliveroo’s OCG and groceries services in response
to Coronavirus (COVID-19) are set out above. While both Parties have
expanded their services in recent months, in our view these changes are
broadly in line with wider developments in the OCG services market in recent
months.
8.284 A groceries retailer submitted that the steps which the Parties had taken in
recent months to expand their OCG services made it clear that they were
‘actual, direct and close competitors in OCG/same day grocery delivery’.
8.285 However, we note that the changes to Amazon’s service which a groceries
retailer highlighted do not relate to on-demand delivery (and indeed Amazon
has cancelled its [] offer). While Deliveroo has launched partnerships with a
number of additional grocery providers, we understand these to be similar in
nature to its existing grocery partnerships ie with a limited range of items.
Deliveroo is also offering a Deliveroo Essentials service independently of
grocery partners, however the service is limited to a small range of items.
8.286 Company D submitted that the combination of Amazon and Deliveroo would
‘place the combined Amazon/Deliveroo entity in a unique position of power’,
and that grocery retailers’ ability to compete will be restricted by their limited
choice of last-mile delivery solutions post-Merger.
456
456
Company D response to the Revised Provisional Findings, 10 July 2020, paragraph 2(c).
245
8.287 However, we note that several alternative last-mile delivery competitors
(including Stuart and Uber Eats) will remain in the market as potential
partners for grocers.
8.288 In our view, there is scope for a significant degree of competitive interaction
between Amazon and Deliveroo as their OCG services evolve, but they are
also likely to face competition from other providers who are well placed to
compete in the market. On balance we do not consider there is sufficient
evidence that Amazon and Deliveroo will have a decisive advantage over
other market participants in the medium to longer-term.
Amazon discouraging Deliveroo from competing against Amazon in OCG
8.289 Turning to our first theory of harm, if Amazon decided to expand its presence
in OCG (including on-demand delivery), or if it was concerned about
Deliveroo’s expansion, it could potentially seek to use its material influence
over Deliveroo to discourage Deliveroo from also expanding its own OCG
offer (which would otherwise be in competition with Amazon). For instance,
Amazon might use its influence to seek to discourage additional investment
by Deliveroo in OCG and/or to prioritise other areas for expansion.
Amazon’s ability to discourage Deliveroo from expanding its OCG offer in
competition with Amazon
8.290 The Transaction would provide Amazon with a 16% shareholding in
Deliveroo, a board seat and certain other rights. Whilst we have concluded
the Transaction is more likely than not to provide Amazon with the ability to
exercise material influence over the policy of Deliveroo, this is not the same
as an ability to control that policy. In particular, it does not amount to an ability
to drive policy in a direction that other shareholders, management or the
board object to.
8.291 As such, we cannot assume that Amazon will be able to drive policy in a
direction that would lead to a substantial lessening of competition if that would
lead to Deliveroo foregoing a compelling commercial opportunity for
Deliveroo. This could make it more difficult for Amazon to discourage
Deliveroo from expanding its OCG business, especially if Deliveroo saw a
strong growth opportunity for OCG.
Effect of Amazon discouraging Deliveroo from expanding its OCG offering
8.292 To the extent that Amazon may be able to influence aspects of Deliveroo’s
strategy in ways that would reduce competitive pressure on Amazon, we also
need to assess whether this would result in an SLC.
246
8.293 To reach that conclusion, first, we need to assess whether without Amazon’s
influence Deliveroo would be likely to expand its OCG offering to compete
more directly with Amazon. As discussed above, Deliveroo’s internal
documents identify groceries as an important opportunity. Deliveroo has
submitted to us that it sees groceries as a ‘bolt-on’ to its restaurant service.
However, we note that it has continued to expand its OCG offer in recent
months and has launched a number of new partnerships, in addition to its own
Essentials service.
8.294 While Deliveroo is likely to expand its OCG offering, it is not clear that this
expansion will bring it into closer competition with Amazon. While Deliveroo
has recently expanded its services, launching partnerships with a number of
additional grocery providers, we understand these to be similar in nature to its
existing grocery partnerships, ie with a limited range of items. Deliveroo is
also offering a Deliveroo Essentials service independently of grocery partners,
however the service is limited to a small range of items. As yet, Deliveroo
does not offer a comparable range to that of Amazon from any grocery
provider, and its logistics network consists of bicycles and mopeds which do
not enable it to deliver larger baskets.
8.295 Second, as regards competition in the market generally, we need to assess
whether Amazon and Deliveroo would emerge as strong players in the
provision of OCG services, and act as an important competitive constraint on
one another, while other market participants would not expand to a sufficient
extent to ensure effective competition.
8.296 As set out in paragraphs 8.260 to 8.264, each of the Parties has advantages
which could mean they are well-placed to compete in OCG services,
although, each also has disadvantages relative to certain other market
participants. Further, as set out in paragraphs 8.277 to 8.280, almost all larger
players in the categories of online delivered groceries, online restaurant
delivery, convenience stores and traditional groceries are currently active in
trialling or offering OCG services with on-demand delivery of between 30-
minutes and two-hours. Most market participants share an intention to expand
their existing offers, and a number are currently in a process of relatively rapid
expansion. As a result, while the future development of the market remains
uncertain, we consider that on the basis of current evidence other market
participants appear well-placed to compete in OCG provision.
8.297 Considering these issues in the round, our view is that the Transaction is
unlikely to lead to an SLC through Amazon discouraging Deliveroo from
competing in online convenience grocery services.
247
Amazon competing less aggressively in OCG services
8.298 We have considered whether Amazon may compete less aggressively against
Deliveroo in the supply of OCG services because of its minority (16%)
shareholding in Deliveroo.
457
In the counterfactual if Amazon increased its
prices or reduced its service quality (eg to save costs), it would expect some
of its customers to switch to other OCG providers, including Deliveroo. But
following the Transaction, if Amazon increased its prices and some customers
responded by switching to Deliveroo, Amazon would have a 16% share of any
increase in Deliveroo’s profits from these customers. As a result, Amazon
may have an incentive to set higher prices following the Transaction than in
the counterfactual.
8.299 Conversely, if Deliveroo faced aggressive competition from Amazon this could
reduce its profitability, which would in turn reduce the profitability of Amazon’s
16% shareholding in Deliveroo. This possibility could lead Amazon to avoid
competing directly against Deliveroo. For example, if Amazon were
developing its own point-to-point delivery network it could potentially decide
not to do so in areas where Deliveroo was present, out of a concern that this
would make it more difficult for Deliveroo to achieve economies across its
courier network or on other fixed costs such as advertising or IT development.
8.300 We have considered both (a) the effect of the Transaction on Amazon’s
incentive to compete in the supply of OCG services; and (b) the effect of
Amazon competing less aggressively in the supply of OCG services.
Effect of the Transaction on Amazon’s incentive to compete aggressively in the
supply of OCG services
8.301 Horizontal unilateral effects of a merger arise when it becomes less costly for
a merged firm to raise its prices (or reduce quality), because it can recoup the
profit from recaptured sales from customers who would have switched to the
other merged firm.
458
8.302 As a minority shareholder in Deliveroo, Amazon benefits from 16% of
Deliveroo’s profits. As such, if Amazon loses a sale to Deliveroo, it retains
16% of the profits Deliveroo earns from the sale, while if Amazon wins a sale
from Deliveroo, it also loses 16% of the profits that Deliveroo would have
457
For the reasons set out in paragraph 7.25, we consider that we should assess the shareholding Amazon will
acquire through the current Transaction and not any possible larger shareholding it may acquire in future.
458
MAGs, paragraph 5.4.8.
248
earned from that sale. This dynamic will tend to reduce Amazon’s incentive to
compete aggressively.
8.303 The question for us is whether a 16% shareholding reduces Amazon’s
incentive sufficiently that it would prevent Amazon from making investments it
otherwise would have made to improve its OCG offering.
8.304 The magnitude of these effects depends on the future closeness of
competition between the two Parties (in the counterfactual), and competition
from other OCG providers (ie the effect would be weaker if most customers
who switched from Amazon switched to other providers rather than
Deliveroo).
8.305 In our assessment of online restaurant platforms (paragraphs 7.35 to 7.37) we
discuss an analysis submitted by the Parties which finds that the acquisition of
a 16% holding does not lead to material horizonal price effects. This analysis
is in the context of a four-to-three merger in online restaurant platforms, and
some specific modelling assumptions. However, in broad terms we would
expect the 16% holding arising from the Transaction to produce a weaker
price effect on OCG services than might be expected from a full acquisition.
Effect of Amazon competing less aggressively in OCG provision
8.306 Even if we had concluded that as a result of the Transaction Amazon would
have an incentive to compete less aggressively in OCG, we also need to
assess whether this would result in SLC in the OCG market.
8.307 Almost all larger players in the categories of online delivered groceries, online
restaurant delivery, convenience stores and traditional groceries are currently
active in trialling or offering OCG services. As discussed above, market
participants share an intention to expand their existing offers, and a number of
OCG services are currently in a process of relatively rapid expansion. As a
result, while the future development of the market remains uncertain, we
consider that on the basis of current evidence other market participants
appear well-placed to compete in OCG provision.
8.308 We consider that this will further weaken any price effect arising from the
Transaction because (a) if Amazon increase its prices, customers may switch
to providers other than Deliveroo, and (b) even if Amazon were to compete
less aggressively against Deliveroo, Deliveroo would still face competition
from these other providers. Considering these issues in the round, our view is
that the Transaction is unlikely to lead to an SLC arising from Amazon
avoiding direct competition against Deliveroo in the provision of OCG.
249
Amazon relying on Deliveroo for its presence in OCG
8.309 Finally, we consider whether the Transaction would harm competition by
reducing the incentive for Amazon to invest in competing more effectively as
an OCG provider, because the proposed acquisition either secures Amazon
an option to acquire Deliveroo or is a first step towards a full acquisition.
8.310 The future development of the OCG market is uncertain. One possibility is
that it will expand rapidly with an increasing focus on a fast (eg one hour)
service with widespread coverage. Given the overall scale of the grocery
sector (see paragraphs 8.8 and 8.9, and Figure 8.1) OCG could reach a very
material scale by winning a relatively small proportion of shopping missions.
8.311 If the market evolved towards faster delivery and Amazon did not have a
widely available one-hour or two-hour service, Amazon could miss out on a
significant opportunity for growth and profits (including flywheel benefits). In
addition, it would face the risks of:
(a) Faster OCG providers winning business from Amazon Prime Now, which
is currently differentiated from the online offers of traditional grocers by
offering approximately [] delivery rather than slower same-day or next-
day delivery.
(b) Rivals using OCG to develop a base of regular customers and a point-to-
point delivery network, from which they could potentially expand to other
areas (eg e-commerce) in competition with Amazon.
8.312 Absent the proposed Transaction, Amazon may therefore have a strong
incentive to invest in developing point-to-point logistics capabilities or by
otherwise improving the delivery speeds it can offer or other aspects of its
service to ensure that it can remain competitive if faster delivery increases in
importance.
8.313 Internal documents from Amazon and other Deliveroo shareholders indicate
that they see the Transaction as [], rather than a purely financial
investment:
(a) An internal email [].
(b) []
(c) []. [].
459
459
[]
250
(d) [] commented on Amazon’s intentions regarding its investment in
Deliveroo: [].
(e) [] noted, []. It also commented that ‘[]’.
8.314 In summary, the evidence indicates that [].
8.315 If Amazon sees the Transaction as giving it a plan or an option for achieving a
stronger presence in OCG provision (ie doing so via Deliveroo), it may be less
likely to have an incentive to improve its own OCG service.
8.316 In our assessment of online restaurant platforms (paragraph 7.77), we have
reached a view on the balance of probabilities that Transaction will not give
rise to an SLC by changing Amazon’s incentives to re-enter that market. Our
view on this point is informed by:
(a) the Parties’ argument that, if Amazon was presented with a compelling
opportunity to re-enter the market, then having a 16% investment in
Deliveroo may not materially alter Amazon’s incentive to undertake such a
strategy;
(b) Amazon’s argument that in order to realise the key benefits of offering
restaurant delivery it would need further integration than a minority
investment;
(c) Evidence to suggest that Amazon may have an interest in pursuing
multiple entry routes into supplying online restaurant platforms; and
(d) Our view Amazon has continued to pursue the development of an online
restaurant platform and if this were to result in a successful restaurant
platform, we would expect this to be rolled out internationally.
8.317 Some of these points are also relevant to the assessment of the present
theory of harm. However, there are differences between the two questions, in
particular in that (i) the online restaurant platform is an established and large
market with a proven business case, whereas OCG is a nascent market, and
(ii) whether Amazon will invest in competing more effectively in a market
where it is already present, building on its existing grocery/OCG provision, is
different from the question of de novo entry into online restaurant platforms.
8.318 In particular, it may be the case that if Amazon were presented with a
compelling opportunity to develop its OCG offer (such as by building a point-
to-point network) then a 16% investment in Deliveroo may not affect its
incentives to do so. On the other hand, since OCG is a nascent market which
does not yet have an established business case, Amazon may see faster
251
grocery delivery as a defensive necessity rather than a compelling
opportunity.
460
In which case it could potentially see its holding in Deliveroo as
providing it with a sufficient ‘[]’ (see paragraph 8.313(c) above) if market
developments towards faster delivery threaten its own grocery plans.
8.319 In considering whether the Transaction may reduce Amazon’s incentives to
improve its own OCG services, we note that:
(a) The fact that, in our assessment, the most likely counterfactual involves
Amazon entering the online restaurants platform in future (and the
Transaction would not affect its incentives to do so) provides a potential
route for Amazon to improve its OCG services, namely by using the same
logistics network for its online restaurants delivery and for some or all of
its OCG services. We note that both logistics-enabled online restaurants
platforms (Deliveroo and Uber Eats) offer OCG services.
461
(b) Amazon may in any case prefer to develop its own OCG service given
that in doing so it could potentially build on its experience in developing
such a service in the US, and [].
(c) While the Transaction potentially gives Amazon a route to acquire
Deliveroo, Amazon would face a significant cost in making such an
acquisition – particularly if Deliveroo’s OCG business increases in value.
It would also face risks eg of being outbid by another buyer.
8.320 Again, this issue needs to be assessed in the context of wider competition in
the OCG services market, and we note that almost all larger players in the
categories of online delivered groceries, online restaurant delivery,
convenience stores and traditional groceries are currently active in trialling or
offering OCG services. Most market participants share an intention to expand
their existing offers, and a number of OCG services are currently in a process
of relatively rapid expansion. As a result, while the future development of the
market remains uncertain, we consider that, on the basis of current evidence,
other market participants appear well-placed to compete in OCG provision.
8.321 In this context, if the Transaction were to reduce Amazon’s incentives to
improve its own OCG services, improvements by its competitors of their own
OCG services could nevertheless lead to effective competition. In addition, we
note that if the market grows substantially this makes it more likely that it will
460
As noted above, it has described its [] as a ‘necessary strategic move’ in light of competitor offers.
461
As does Just Eat, using the same logistics provider for both restaurant and OCG delivery.
252
be seen as a compelling commercial opportunity, both by Amazon and its
competitors.
8.322 Considering these issues in the round, our view is that the Transaction is not
likely to lead to a SLC by removing the strategic benefit to Amazon of
developing its own OCG service.
9. Additional theories of harm
9.1 Our Issues Statement
462
set out that we would consider theories of harm
relating to horizontal unilateral effects in online restaurant platforms and
online convenience grocery delivery in the UK, and our assessments of these
issues are in Chapters 5 and 6. We did not receive any responses to our
Issues Statement. However, later in our inquiry, some third parties have
raised additional competition concerns about the impact of the Transaction.
We consider these concerns in this chapter.
9.2 We considered the following concerns in our Revised Provisional Findings,
and these are discussed in turn below.
(a) That the Transaction will allow Amazon to bundle Amazon Prime and
Deliveroo’s services, with a negative impact on other providers.
(b) That Amazon will prevent other OCG providers from having access to
Deliveroo’s logistics network.
9.3 In response to our Revised Provisional Findings, Company E submitted that
start-ups and smaller operators already face challenges in competing against
Deliveroo, and that if Deliveroo is supported by Amazon it will be a dominant
player, with an increase in market share and margins.
463
As noted in
paragraph 7.73, we consider that Deliveroo faces strong competition from
Just Eat and Uber Eats at the national level.
Bundling of Deliveroo in Amazon Prime
9.4 In response to our April Provisional Findings,
464
two third parties expressed
concerns relating to the possible bundling of Deliveroo within Amazon Prime.
462
Issues Statement, 28 January 2020, paragraphs 24 and 25.
463
Company E response to the Revised Provisional Findings, 1 July 2020.
464
April Provisional Findings, 16 April 2020.
253
Third party views
9.5 Company D submitted that post-Transaction, Amazon has an incentive to
promote its own groceries business, and potentially its wider business, by
offering Deliveroo’s service as part of Amazon Prime.
465
It said that the
Transaction would allow Amazon to offer its Amazon Prime customers cheap
or free delivery, or more preferential delivery timings, on Deliveroo.
Company D submitted that this could: (a) drive additional customer traffic
through Deliveroo’s app at the expense of other online restaurant
platforms;
466
and/or (b) drive additional traffic to Amazon’s on-demand
grocery offering at the expense of other grocery retailers (ie because the
inclusion of preferential access to Deliveroo within the Amazon Prime offer
would increase demand for Amazon Prime and Amazon Groceries). It
submitted that, given that Amazon Prime already covered a substantial
proportion of UK households, any impact of a strategy of this kind would have
a considerable negative effect on other players.
467
9.6 Company D further submitted that Amazon stands to benefit financially from
bundling by ‘pushing its own groceries via Deliveroo’ and by enhancing its
‘flywheel’ benefits; ‘Amazon is one of the wealthiest companies in the world,
and (by its own admission) tends to adopt a “test and learn” approach. There
is no reason to conclude that it wouldn’t look to engage in a (potentially
predatory) strategy of this kind’.
468
9.7 Domino’s submitted that the Transaction could raise concerns due to
(a) Amazon offering Deliveroo’s services to its users at no or low cost, such
that rivals such as Domino’s could not compete,
469
and (b) offering free
delivery of Deliveroo orders ‘for Amazon Prime or Amazon Fresh
members’,
470
tying/bundling Amazon Prime and Deliveroo Plus subscriptions,
or offering Amazon Whole Foods groceries at a preferred tariff.
9.8 Domino’s noted ‘that the CMA concluded in phase 1 that there was insufficient
incentive for the Parties to offer bundled products since the bundle would
need to be offered at “a deep discount” compared to the standalone
465
Company D response to the April Provisional Findings, 11 May 2020, paragraph 29(a).
466
Company D similarly submitted in response to the Revised Provisional Findings that bundling would ‘grant
Deliveroo preferential access to the extensive Amazon Prime customer base’ (Company D response to the
Revised Provisional Findings, 10 July 2020, paragraph 2.c.(iv)(5)).
467
Company D response to the April Provisional Findings, 11 May 2020, Annex A.
468
Company D response to the Revised Provisional Findings, 10 July 2020, paragraph 2.c.(iv).
469
Domino’s refers to this as ‘targeted advertising’ (Domino’s response to the April Provisional Findings, 7 May
2020, paragraph 100), however as described it seems to be a matter not of advertising, but of offering Deliveroo’s
service at a subsidised price. In addition, Domino’s did not explain how this offer would be targeted, or how
targeting would in itself cause a problem.
470
This appears to be similar to the ‘targeted advertising’ concern.
254
subscriptions. However, Domino’s considers that the long-term benefit derived
by Amazon would likely be worth an initial deep discount’.
471
Domino’s
provided no further elaboration of this point, and did not refer to any evidence
in support of its view.
472
9.9 Company A submitted that Amazon’s incentive to bundle Deliveroo into
Amazon Prime arose from its (Amazon’s) objective of growing the number of
subscribers to Amazon Prime, rather than any profits Amazon stands to gain
from its stake in Deliveroo.
473
9.10 Both Company A and Company D submitted that we should consider how any
benefit Amazon would derive from a bundle (and as a result, Amazon’s
incentives to support a bundled offering) would increase as it increased its
shareholding in Deliveroo.
474
9.11 Finally, Company A submitted that the CMA should consider seeking
commitments from Amazon which would, among other things, limit Amazon’s
shareholding to the level resulting from the initial investment without prior
approval from the CMA.
475
Our assessment
9.12 As set out in the CMA’s Phase 1 Decision,
476
conglomerate effects may arise
in mergers of firms that are active in the supply of goods or services that do
not form part of the same markets but which are nevertheless related in some
way, for example because their products are complements (so that a fall in
the price of one good increases the customer’s demand for another) or
because there are economies of scale in purchasing them (so that customers
buy them together).
9.13 As set out in the CMA’s Merger Assessment Guidelines, a merger between
providers of complementary products may result in a bundle of those products
being offered at a price below their standalone prices, and bundling can also
lower transaction costs to the consumer by offering a ‘one stop shop’.
477
The
471
Domino’s response to the April Provisional Findings, 7 May 2020, paragraph 100.
472
In its response to the Revised Provisional Findings, Domino’s argued that we had failed to consider the
concerns about bundling identified in its response to the April Provisional Findings (Domino’s response to the
Revised Provisional Findings, 10 July 2020, paragraphs 111–112). These concerns were considered alongside
the similar concerns identified by Company D and Company A and have been considered again in reaching this
final decision.
473
Company A response to the Revised Provisional Findings, 10 July 2020, paragraphs 12–13.
474
Company A response to the Revised Provisional Findings, 10 July 2020; Company D response to the Revised
Provisional Findings, 10 July 2020.
475
Company A response to the April Provisional Findings, 7 May 2020, paragraph 13.
476
Phase 1 Decision, paragraphs 299–364.
477
MAGs, paragraphs 5.7.17–5.7.18.
255
CMA’s approach to assessing conglomerate theories of harm is to analyse
(a) the ability of the merged entity to foreclose competitors, (b) its incentive to
do so, and (c) the overall effect of the strategy on competition.
478
9.14 In phase 1, the CMA considered: whether the Parties could bundle Amazon
Prime and Deliveroo Plus by offering Deliveroo Plus to Prime members at a
discounted price, or alternatively by offering Deliveroo Plus as a free benefit to
all Prime customers (likely with a price increase across the Prime subscriber
base); and if so whether this could foreclose (or prevent) either other food
delivery companies or other suppliers of subscription services from competing
with them.
9.15 At phase 1, the CMA found, on a realistic prospect basis, that the Parties may
have the ability to use this strategy to foreclose Deliveroo’s competitors.
However, even though the Parties would have the ability to bundle Deliveroo
Plus and Prime, the CMA did not find that they would have the incentive to do
so.
9.16 The CMA concluded that, for a bundle of Deliveroo Plus and Amazon Prime to
be successful in attracting a large volume of customers, the Parties would
need to offer the bundled product to those customers at a substantial discount
to the price of Amazon Prime and Deliveroo Plus on a standalone basis. The
level of Amazon’s shareholding in Deliveroo would limit its incentive to fund a
large discount, as Amazon would only gain a small part of the benefit from
each additional customer won by Deliveroo.
9.17 While it is possible that Amazon and Deliveroo will at some point offer a
bundle of their services, this could only have the foreclosure effect considered
at phase 1 if the bundle were offered at a sufficient discount to attract a large
volume of customers. Accordingly, the CMA concluded that the Transaction
did not give rise to a realistic prospect of an SLC as a result of conglomerate
effects.
9.18 The Phase 1 Decision recognised that Amazon may be able to offset some of
the costs of subsidising a discount on Deliveroo Plus via the combined
offering through ‘flywheel benefits’.
479
More generally, the Phase 1 Decision
recognised that there may be benefits to Amazon of bundling with Deliveroo.
However, despite these potential benefits, the CMA considered that the
Transaction would not lead to a sufficient change in Amazon’s incentives to
make a bundling foreclosure strategy profitable.
478
MAGs, paragraph 5.6.6.
479
Phase 1 Decision, paragraph 346.
256
9.19 Furthermore, as regards whether the long-term benefit of bundling to Amazon
would likely be worth an initial deep discount, the CMA specifically considered
the long run incentives of the Parties when reaching its view, at phase 1, that
the incentive arising from the Transaction to engage in this strategy was
insufficient to confer a realistic prospect of an SLC:
‘The Merger implies Amazon will internalise only a relatively low
level (based on its shareholding) of Deliveroo’s profit margins,
which would be insufficient to make a profit recoupment strategy
in the long run profitable, and further insufficient to reduce the
difficulties, costs and risks that the Parties face in reaching
agreement on a strategy that would provide for the deep
discounts needed to foreclose Deliveroo’s competitors’.
480
9.20 In addition, the CMA noted that Amazon’s shareholding in Deliveroo did not
increase Deliveroo’s incentive to discount Deliveroo Plus through Amazon
Prime.
9.21 The CMA therefore concluded at phase 1 that the Transaction did not give
rise to a realistic prospect of an SLC as a result of conglomerate effects (ie
bundling of Deliveroo Plus and Amazon Prime) in the supply of online food
delivery platforms in the UK.
9.22 We consider that the concerns raised by Company D and Company A in
relation to bundling of Deliveroo and Amazon Prime are, in substance, the
same as those which the CMA considered in its Phase 1 Decision, in which
the CMA concluded that the Transaction did not give rise to a realistic
prospect of an SLC as a result of conglomerate effects.
481
9.23 We have not seen any new evidence (and, in particular, Company D and
Company A have not submitted any evidence to support their submissions)
that would lead us to reopen the phase 1 conclusion on this point.
482
9.24 Regarding the point raised by Company A and Company D about the effect
that an increase in Amazon’s shareholding would have on its incentive to
engage in bundling, we note that the purpose of this inquiry is to consider the
Transaction’s impact on competition. As explained in paragraph 7.25, we do
not agree that it is incumbent on the CMA to consider the ‘maximum’ version
480
Phase 1 Decision, paragraph 361(b).
481
Phase 1 Decision, 27 December 2019.
482
Under the UK’s two-phase merger control regime, the CMA is required to apply different evidentiary
thresholds at phase 1 and phase 2 when answering the statutory questions in the Act. At phase 1 the CMA
applies a ‘realistic prospect’ threshold whereas at phase 2 it applies a (higher) ‘balance of probabilities’ threshold.
See MAGs, paragraph 2.2.
257
of material influence that might arise. Moreover, increases in shareholding
that lead to a change in the quality of Amazon’s control over Deliveroo may be
subject to the CMA’s merger control jurisdiction.
9.25 Finally, Company A submitted that a remedy should be imposed on the
Parties to address concerns about potential bundling.
483
We are required to
consider whether the Transaction is more likely than not to result in an SLC.
As noted in paragraph 1.3 the CMA only has the power to impose remedies
as part of a merger review where it concludes that a transaction will result in
an SLC. We do not otherwise have a discretion to take actions such as
seeking commitments from Amazon (as suggested by Company A).
Additional concerns raised in responses to the Revised Provisional Findings
9.26 Domino’s submitted that the Revised Provisional Findings failed to address
concerns about an ‘ecosystem’ theory of harm.
484
Domino’s said that this
theory of harm had been set out by Cremer et al
485
and discussed in the
CMA’s Digital Advertising Market Report.
486
Domino’s submitted that Amazon
has a dominant ecosystem and identified concerns about ‘opportunities for
cross-selling between Amazon and Deliveroo’s online restaurant platform
business the Transaction would create, either by way of targeted advertising
or tying/bundling’.
487
9.27 We have not considered it necessary to reach a conclusion as to whether
Amazon is dominant in any segment in the context of this inquiry. We
recognise that various academic studies and reports have suggested that
Amazon may hold a dominant position in one or more segments.
9.28 We recognise that Amazon could be considered to have an ‘ecosystem’ of
complementary products and services. The concerns raised by Domino’s with
respect to Amazon’s ecosystem appear to relate primarily to potential tying or
bundling, and those concerns are addressed above. In addition to the tying or
bundling concerns, concerns relating to Amazon’s position in other segments
have informed a number of the theories of harm which we have considered:
(a) As set out in Chapter 8, we have assessed whether the Transaction may
give rise to horizontal unilateral effects in the supply of OCG services. In
doing so we considered (paragraph 8.260) a number of advantages which
483
Company A response to the Revised Provisional Findings, 10 July 2020, paragraphs 18–20.
484
Domino’s response to the Revised Provisional Findings (10 July 2020). See also Domino’s response to the
April Provisional Findings (7 May 2020).
485
Cremer et al, 20 May 2019, Competition Policy for the Digital Era.
486
Online platforms and digital advertising market study, Final report, 1 July 2020. The report describes certain
companies as having built large ‘ecosystems’ of complementary products and services around their core service.
487
Domino’s response to the Revised Provisional Findings, 10 July 2020, paragraph 26.
258
may put Amazon in a strong position to compete as the market evolves,
including its base of Prime customers to whom it could promote OCG
services, the logistics model it uses for Prime Now, and its potential to
attract more customers to Prime via its OCG offer and thus achieve
‘flywheel’ benefits which are not available to other competitors.
(b) At phase 1, the CMA considered:
488
(i) Whether, absent the Transaction, Deliveroo might have expanded to
start delivering more non-food items in competition with parts of
Amazon’s broader offering, and if so whether the Transaction would
give rise to a realistic prospect of an SLC as a result of horizontal
unilateral effects in the supply of logistics-enabled e-commerce
marketplaces in the UK. The CMA did not find a realistic prospect of
an SLC on this basis, in particular absent evidence that Deliveroo
planned to expand into non-food categories.
(ii) Whether the Transaction might lessen competition by increasing the
Parties' incentives to share data with each other which they could use
to engage in behavioural discrimination in the supply of online food
platforms or logistics-enabled e-commerce marketplaces in the UK.
9.29 Company D submitted that we should have carried out ‘the detailed analysis
expected of a phase 2 investigation’ rather than relying on a phase 1
analysis.
489
9.30 We note that the Phase 1 Decision concluded that Amazon’s 16%
shareholding in Deliveroo would be insufficient to affect Amazon or
Deliveroo’s incentives to engage in a bundling strategy.
490
The factors that led
to that conclusion remain unchanged. As set out in our Issues Statement,
491
we have focused our inquiry on the areas in which the CMA had found a
realistic prospect of an SLC at phase 1. As also set out in that Issues
Statement, we considered that we could assess additional issues (including
the theories of harm dismissed at phase 1) if we found new evidence
suggesting they could give rise to concerns. We invited interested parties to
respond to our Issues Statement and did not receive any responses
suggesting either that there was evidence supporting other concerns, or that
we should investigate any other concerns. No additional evidence supporting
other concerns has been received by the CMA subsequently.
488
Phase 1 Decision, paragraph 15.
489
Company D response to the Revised Provisional Findings, 10 July 2020, paragraph 2.c.(iv).
490
Phase 1 Decision, paragraphs 359–362.
491
Issues Statement, 28 January 2020, paragraphs 7 and 8.
259
Preventing third party access to Deliveroo’s delivery network
9.31 Two third parties submitted that following the Transaction Amazon could
prevent rival grocery suppliers from accessing Deliveroo’s logistics network at
competitive prices.
9.32 The concern is one of partial input foreclosure, arising from the ‘vertical’
combination between the provision of an upstream input (Deliveroo’s logistics
network) and a downstream product (OCG services). In such cases, the
theory of harm to be assessed is that the merged firm could increase the price
it charges for the input (on-demand delivery) to rival OCG providers which, in
turn, could make it harder for rival OCG providers to compete by increasing
their costs, making them less competitive, and lessening downstream
competition.
492
9.33 As with conglomerate effects, the CMA’s approach to assessing partial input
foreclosures is to analyse (a) the ability of the merged entity to foreclose
competitors, (b) the incentive of it to do so, and (c) the overall effect of the
strategy on competition.
493
Third party views
9.34 In response to our April Provisional Findings, Company D submitted that:
(a) The Transaction will give Amazon ‘the ability (via its material influence)
and incentive to impede the ability of other players to use Deliveroo to
offer OCG services to consumers at reasonable prices’.
(b) ‘Once Amazon secures exclusive use of Deliveroo for its own benefit, it
will be extremely challenging for other players to access customers in the
OCG market through the remaining two point-to-point delivery
networks’.
494
9.35 Company D previously submitted that:
(a) Amazon would control Deliveroo as a result of the Transaction;
(b) Amazon will have no incentive to afford traditional grocery retailers the
opportunity to access/use Deliveroo as a key route to market for online
convenience groceries, either at all or on fair and reasonable terms;
492
See MAGs, paragraph 5.6.9.
493
See MAGs, paragraph 5.6.6.
494
Company D response to the April Provisional Findings, 11 May 2020, paragraph 29(a)–(b).
260
(c) there are limited alternative routes to market for online convenience
grocery (only Uber Eats, Just Eat and a handful of third-party logistics
companies, none of which have the requisite scale to grow at pace);
(d) several traditional grocery retailers have taken steps to enter or grow
within this space, often using the services of Deliveroo to do so. The
Transaction would stifle this emerging competition and give Amazon an
unfair advantage;
(e) a first-mover advantage is key to entrenching a strong position in OCG
services; and
(f) the revenues Amazon may stand to earn via its minority stake in allowing
traditional grocery retailers to sell groceries through Deliveroo will be
significantly lower than the benefits it could reap by leveraging Deliveroo
to its exclusive advantage.
9.36 A groceries retailer told us that the biggest challenge of using online delivery
platforms is commission and making it economical to use them. Online
delivery platforms own rider networks which can deliver more effectively in
terms of scope and scale than Stuart, and it would like access to Deliveroo’s
network but not at the cost Deliveroo wants to charge.
9.37 Company D submitted that it would be considerably easier and financially
more attractive for Deliveroo to deal with a single supplier of groceries (ie
Amazon) with an extensive, active and loyal customer base (ie Amazon
Prime), rather than the range of retailers with which it currently partners. In
assessing the attractiveness of Amazon as a groceries partner, it is irrelevant
to consider Amazon’s physical presence in groceries. Amazon has
considerable access to groceries and extensive ambitions to grow in this
space.
9.38 Company D also said that ‘an exclusive arrangement between Amazon and
Deliveroo would leave only Just Eat and Uber Eats as alternative routes to
market – neither of which currently has a strong presence in OCG’. Third
party couriers (such as Stuart) are rapidly reaching saturation levels in terms
of the scale and scope of the services they can render to grocery retailers.
495
9.39 Company A submitted that:
‘Preventing Deliveroo from engaging in exclusive commercial
arrangements with Amazon would help avoid the situation where
495
Company D response to the Revised Provisional Findings, 10 July 2020, paragraph 2.c.(v).
261
other online convenience grocery suppliers are deprived of one of
the most credible delivery routes to market.
As the CMA held at phase 1, supermarkets do not have the
necessary logistical capabilities to offer online convenience
groceries and are reliant upon businesses such as Deliveroo,
Uber Eats, or third party logistics providers like Stuart in order to
offer ultrafast delivery services.
A positive obligation requiring Deliveroo to engage with partners
other than Amazon would prevent Deliveroo from sourcing all of
its convenience groceries from Amazon and would ensure that
other grocery retailers continue to have a route to market. This
would help prevent the nascent online convenience groceries
market segment from developing into one dominated by Amazon
and would ensure that the Deliveroo delivery platform remains
open for other competitors to use’.
496
9.40 In addition, one retailer told us it believed that Deliveroo and Uber Eats could
increasingly become the instinctive starting place for consumers on an online
convenience grocery mission, with the result that grocery brands wishing to
play in this space may have little choice but to list on Deliveroo or Uber Eats.
It added that, without significant resources, it would be difficult for grocery
brands to reach sufficient scale in order to deliver sustainable economics and
compete with third party delivery platforms, particularly if they are financially
backed by the likes of Amazon, and the result would be a stifling of
competition. On the other hand, Co-op told us that it expected an increase in
competition as the main competitors have realised the demand for online
convenience groceries and this will likely form a bigger part of their future
strategies.
Our assessment
9.41 As discussed further below, this concern is in some respects related to our
horizontal unilateral effects theories of harm, which was considered at
phase 1 and on which our view is set out in in Chapter 6. However, the CMA’s
Phase 1 Decision did not specifically consider a vertical effects theory of
harm, arising from concerns about third-party access to Deliveroo’s delivery
network.
497
496
Company A response to the April Provisional Findings, 7 May 2020, paragraph 13(c)–(d).
497
Phase 1 Decision, 27 December 2019.
262
9.42 As noted above, our approach is to consider the ability of the merged entity to
foreclose competitors, its incentive to do so, and the overall effect of the
strategy on competition. We begin by considering ability and incentive in the
round, in relation to Amazon’s ability to influence Deliveroo to foreclose
competitors.
Amazon’s influence on Deliveroo
9.43 As set out in paragraphs 8.290 and 8.291, we consider that the Transaction is
more likely than not to provide Amazon with the ability to exercise material
influence over the commercial policy of Deliveroo (see further the April
Provisional Findings
498
). This is not the same as an ability to control that
policy. In particular, it does not amount to an ability to drive policy in a specific
direction against the objections of other shareholders, management or the
board.
9.44 Company D has not explained why it is illogical to distinguish between
‘material influence’ and ‘control’.
499
As such, we do not assume that Amazon
will be able to drive policy in a direction that would lead to Deliveroo foregoing
a compelling commercial opportunity.
9.45 The evidence to date indicates that Deliveroo has seen, and continues to see
a compelling commercial opportunity in partnering with a wide range of
grocery suppliers:
(a) A June 2019 Deliveroo strategy document identified grocery as []. To
the extent that Deliveroo continues to pursue either of these opportunities,
they will give third party grocery suppliers the opportunity to access
Deliveroo’s logistics network for grocery delivery.
(b) Deliveroo has pursued this opportunity by delivering groceries through its
app from a range of grocery providers, including Co-op, Nisa, BP and
Shell.
(c) Since the beginning of March 2020, Deliveroo has begun delivering from
140 Morrisons stores (from a single-store trial), launched a trial with Aldi
and launched partnerships with BP (M&S food range), M&S (Standalone),
Holland & Barratt, and McColls.
9.46 From Deliveroo’s perspective, the attractiveness of an exclusive or
preferential supply arrangement with Amazon would depend on the value to
498
April Provisional Findings, 16 April 2020.
499
This distinction is made in the MAGs, paragraph 3.2.5 and the differences between material influence and a
controlling interest are highlighted in the CMA’s Jurisdictional Guidance (see CMA2).
263
Deliveroo of such an arrangement relative to the partnerships with other
grocery providers which it would forego in pursuing such an exclusive
arrangement. At present, Amazon has a relatively limited physical presence in
groceries, consisting of [] fulfilment centres and seven Whole Foods stores.
On the other hand, these tend to be located in high-value local markets such
as west London, where Deliveroo also has a substantial network presence,
and Amazon currently delivers groceries in partnership with Morrisons. In
addition, we note (see paragraphs 8.108 to 8.111) that Amazon plans to
[].
500
Although an exclusive arrangement with Amazon might have some
advantages for Deliveroo (principally the simplicity of dealing with a single
supplier), it would also have material disadvantages. Under an exclusive
arrangement, Amazon’s physical presence in groceries would significantly
limit the scale and geographic coverage Deliveroo could achieve for its OCG
offer compared to Deliveroo’s current approach of dealing with a number of
grocers.
9.47 Finally, we note that any attempt by Amazon to restrict third party access to
Deliveroo’s network would likely be to the advantage of Uber Eats and Just
Eat and, as such, if Deliveroo were to engage in such a strategy, it would risk
strengthening its two key competitors in its core market of online restaurant
platforms. We consider that this effect could weaken Deliveroo’s incentives to
pursue a foreclosure strategy.
501
Impact on competition
9.48 Next we consider the likelihood of competitive harm arising if Deliveroo were
to restrict or prevent other OCG providers from accessing its delivery network
or app.
9.49 To date, Deliveroo has largely been focused on impulse/immediate shopping
missions. In this it differs from the nascent OCG offers of online grocery
suppliers and supermarkets which offer a wider range and larger baskets.
Until recently (and in contrast to Company D’s claim in paragraph 9.35(d)
above, these providers have largely contracted a third party (Stuart) for their
grocery delivery rather than Deliveroo or other online restaurant platforms.
OCG providers have noted that the advantage to them of using Stuart is that it
allows them to maintain control over customer relationships. Contrary to the
submission from Company D, we have seen no evidence of ‘saturation’ from
third-party couriers – rather we understand their business models to be
500
While these plans pre-date the Coronavirus (COVID-19) crisis, Amazon has told us that it has not materially
changed its [] plans in response to the crisis.
501
We note that, in principle, even if Deliveroo faced limited competition in OCG it may benefit from competition
between grocers in OCG services, as this could help to expand the market.
264
scalable to new areas/increased demand once they can recruit enough
additional couriers.
9.50 As discussed in Chapter 6 Deliveroo’s offer and capability may evolve as the
market develops, and the expansion of its deal with Morrisons in recent
months may be a step in this direction. On the other hand, such an evolution
will require Deliveroo to address a number of challenges, particularly if it is to
serve a wider range of shopping missions. As an illustration, in late 2019
Sainsbury’s conducted a small trial with Deliveroo for delivery of hot food and
a limited range of groceries. Sainsbury’s reported that the trial resulted in [],
and identified challenges (in technology, managing product substitutions, food
safety, payments processing and labour arrangements) which it would need to
address in order to scale the offer.
9.51 The importance of Deliveroo as a route to market for OCG delivery for third
parties will depend on the future scale of its logistics network, and potentially
the popularity of its app. Deliveroo developed its network and app to serve its
primary market ie online restaurant delivery, and Deliveroo’s future
importance to OCG is likely to be driven to a significant extent by competitive
outcomes in online restaurant delivery. While Deliveroo is a strong player in
this market, it faces competition from Just Eat and Uber Eats. Uber Eats has
an extensive point-to-point delivery network in the UK, while Just Eat is
seeking to develop such a network. Just Eat is the market leader in online
restaurant delivery platforms and has recently been acquired by
Takeaway.com.
502
The impact of these developments on the UK market
remain to be seen, but both Uber Eats and Just Eat potentially offer UK
grocers an alternative route to market. In addition, if Amazon were to enter the
market and compete independently this could also be expected to have an
impact on Deliveroo.
9.52 In our view, the available evidence does not indicate that Deliveroo has
achieved a ‘first mover advantage’ in OCG delivery.
503
Its OCG offer has until
recently been at a relatively nascent stage. It has expanded its offer in
response to the Coronavirus (COVID-19) crisis, as described above. Uber
Eats told us that it had [] grocery/convenience suppliers to its app and had
[] the number of stores on its app since 1 March 2020. As discussed in
paragraph 8.279, a number of grocers have also expanded their OCG
services during the Coronavirus (COVID-19) crisis.
502
See Takeaway.com N.V./Just Eat plc merger inquiry.
503
One respondent to the Revised Provisional Findings suggested that [].
265
9.53 Finally, and as set out in our market definition assessment (paragraphs 5.196
to 5.217, we note that grocery suppliers have options other than Deliveroo for
OCG delivery:
(a) Uber Eats and Just Eat: the future attractiveness of these firms as OCG
delivery partners will depend in large part on their performance in online
restaurant delivery (which is likely to affect their scale and app traffic) and,
in the case of Just Eat, its success in developing an independent logistics
network.
(b) Third-party logistics: using a third-party logistics provider for OCG delivery
can have benefits in that (i) the logistics network may be better suited to
OCG delivery than those of online restaurant platforms, (ii) use of a third
party operator allows the grocer to own the customer relationship.
504
In
practice, a number of the most developed OCG offers are fulfilled by a
third-party logistics provider (Ocado Zoom, Sainsbury’s Chop Chop,
Waitrose Rapid currently only use Stuart, while Co-op uses Stuart and
Ecargo for its Quick Shop offer).
(c) In-house supply: while grocers have identified challenges to developing
their own logistics networks for supply of OCG services, some are already
developing such networks or considering ways to address these
challenges.
Our view
9.54 The possibility of competitive harm arising from the foreclosure concern
outlined above depends on:
(a) whether Amazon will be able to drive policy in a direction that would lead
to Deliveroo foregoing a compelling commercial opportunity; and
(b) whether absent competition between Amazon and Deliveroo other OCG
providers would not ensure effective competition in the market.
9.55 As set out in our assessment of horizontal unilateral effects in the OCG
market our view is that:
(a) we cannot assume that Amazon will be able to drive policy in a direction
that would lead to a SLC if that would lead to Deliveroo foregoing a
504
Although we note this could also be an option from restaurant delivery platforms if they were to offer their
logistics network for ‘white label’ OCG supply.
266
compelling commercial opportunity for Deliveroo (see paragraph 8.291);
and
(b) while the future development of the market remains uncertain, we
consider that on the basis of current evidence other market participants
appear well-placed to compete in OCG provision (see paragraphs 8.277
to 8.280).
9.56 Accordingly, we have not carried out further investigation on this foreclosure
concern.
10. Conclusion on the SLC test
10.1 We have concluded that the anticipated acquisition by Amazon of certain
rights and a 16% minority shareholding in Deliveroo may not be expected to
result in an SLC within a market or markets in the UK for goods and services.
