2) Selection – This isn’t just the number of restaurants
and grocers we have in the neighbourhood. How many
different cuisine types do we have? How many of these
are outlier restaurants/grocers in terms of popularity?
Do we have a wide price range of restaurants? Do
we have exclusive, amazing content? Do we have the
beloved local Chinese restaurant, the halal butcher and
the most popular grocer in the neighbourhood?
3) Consumer experience – It’s not just about how fast the
delivery arrived; this covers the full customer journey.
How accurate was our timing? How easy was it to track
your order? How was the packaging? It’s also about the
quality of personalisation and merchandising in the app.
Do we know, before you do, what you are looking for? Can
restaurants and grocers tell their story effectively and
emotionally in our app? And occasionally, things can go
wrong. How quickly did we resolve the issue?
4) Price – In the end, this is about perception of value, but
a lot of factors influence that. How is our consumer
pricing in terms of delivery fees? What is the price of the
food itself on the platform – and is it marked up over the
restaurant prices? What is the prevalence and relevance of
promotions? Are you part of our subscription programme
Plus, which unlocks free delivery and other rewards?
5) Brand – To me, this is about what Deliveroo stands for
outside of the singular transaction. Does our brand
resonate with people? Are we seen as ethical?
Some of these pillars are easily quantifiable, others not. But
we try to measure them neighbourhood by neighbourhood,
on a standalone basis as well as against our competitors.
This is our scorecard, and it is how we manage the business
and make certain day-to-day operational decisions in a
decentralised manner. For example, general managers can
decide to encourage more riders to work in an area if rider
supply is too low. They can choose to offer a famous local pizza
restaurant an exclusive contract. They can push hyperlocal
in-app discounts, or build a hyperlocal consumer reactivation
campaign to get the flywheel spinning quickly again.
Guided by this report card, we improve over time by grinding
out daily gains, as well as making step-change advances
through long term innovation. Technology is key to both. We
are continuing to improve things we’ve been working on
for years: how long it will take a restaurant to prepare an
order, which rider to assign to collect it, what restaurants
to show to a customer first, how to match rider supply with
demand in real time, which consumer acquisition channels
are most effective. We are also focusing on questions that
have become priorities more recently: how should the user
interface differ for restaurant versus grocery orders, when
should we show upsell items, which orders can be batched,
how best to pick a grocery order, how to build a high quality
advertising platform that brings value to all sides.
These questions are all answered with technology, each
involving teams of data scientists, product designers and
software engineers, to name but a few. They are frequently
hard questions with complex answers. For example, in
the early days our rider assignment algorithm was a
simple ‘greedy solver’ where the closest available rider
would receive the order; now we use deep learning to
predict future network states and advanced optimisation
techniques to decide rider assignment, and we have vastly
more data on which to train our models. This illustrates why
Deliveroo is at its heart a technology company, and why
we’re continuing to invest in our technology team.
Our technology and our operational teams are key to
executing on a hyperlocal basis. This is how we gain market
share in each neighbourhood. This is critical to generating
attractive financial returns. As for any company, overall
scale helps to spread marketing costs and overheads.
But in our business, hyperlocal network effects are more
powerful than overall scale, and network effects come from
hyperlocal market share. Improving and winning local market
share positions yields outsized unit economics, and unit
economics is the key to overall profitability.
Profit pool potential is a function of population density,
affluence, restaurant and grocery partner supply, and our
local market share. Not every neighbourhood is created
equal in terms of potential, but we believe the vast majority
of neighbourhoods in the markets where we operate have
the fundamental demand and supply characteristics to be
profitable. Just how profitable depends in large part on the
strength of our local market position. In the UK, for example,
we believe over 70% of our Gross Transaction Value (GTV*) is
in neighbourhoods where we are number one in terms of
market share. We will aim to increase this percentage across
all our markets, and we will consider exiting neighbourhoods
where we cannot achieve this position.
We operate in a very competitive market, so how do we
maintain durable advantages? Part of it comes down to the
day-to-day execution that I already described – which is
becoming increasingly automated over time. Alongside this
is the combination of long-term vision with the curiosity to
innovate and the willingness to adapt. How do we decide
what new verticals or businesses to enter?
I start by looking through the lens of the three sides of
the marketplace. These are really three core assets in our
business: (1) monthly active consumer (MAC) base; (2) rider
base; and (3) restaurant and grocery partner base. New
verticals or businesses are best if they involve at least
two sides. For instance, Signature, our white label business,
makes sense because we can engage both the partner base
as well as the rider base. Grocery has allowed us to leverage
our MACs as well as our rider base. Editions has allowed us
to utilise all three! Initiatives that only involve one side of the
marketplace can be beneficial, but are less obvious.
Hopefully this provides an understanding of how I think
about the business, and how we make decisions to enter
new areas.
* To supplement performance assessment, Deliveroo uses Alternative
Performance Measures (APMs), which are not defined under IFRS. APMs are
indicated in this document with an asterisk (*); definitions and further
details are provided on page 191.
Strategic report Governance report Financial report
03Annual Report 2021 deliveroo plc