Deliveroo was founded in 2013, and has its headquarters in London. From its origins in the UK, today the company serves markets around the world including France, the United Arab Emirates, Hong Kong, Singapore, and Australia. We discussed with Siddharth Shanker, General Manager of Deliveroo Singapore, about how the company uses its data in order to make sound business decisions that serve the restaurants, consumers and riders they engage.
Data driven decisions
Deliveroo is a company in constant evolution, with changes happening every 6 months to a year. This, according to Siddharth, is “part and parcel of being in a high growth company like this”.
Key to their evaluation and change is having a strong business strategy, which guides how they organized their teams and structure: sometimes centralizing functions and sometimes having them more localized and autonomous.
According to Siddharth, the data that Deliveroo receives across restaurants, consumers and riders is key in underpinning all their business decisions. All decisions made are data driven, and each team is very data proficient, with members who can program SQL and have the necessary statistical capabilities to pool data, understand, analyse and conduct experiments. While the company does have business intelligence teams to handle the most complex pieces, by and large, local teams manage and interpret data independently to make decisions.
Siddharth shared how the setting up of Deliveroo’s Edition Kitchens is a clear example of this approach. Although the kitchen itself is a physical site, a huge amount of data is considered both before and after it’s built. In deciding where the kitchen should be, for example, Deliveroo examines the geography and demographic of the suburb, the demand in the area, what cuisines are being sold, what are the most popular, the price points available, and so on. This is a combination of their own data, as well as external sources:
“Somethings which we might get externally – demographics data, household data and so on, so it’s a combination of those things; how many corporates in a certain area, is there going to be demand from corporates, corporate type meals, which are at slightly different price points. We overlay all this data to make a decision.”
By looking at these consumer preferences and data points they discover what would really do well there, but is missing. Thus, by bringing those in, they will see demand. According to Siddharth, this is the first point in deciding whether its worth investing capital to build an editions centralized kitchen in that area, and which restaurants they reach out to.
Before the site is built, there is already a data driven view on the cuisines and price points, which they use when advising the restaurants. For example VIOS, a branch off from Blu Kouzina, uses the same ingredients, chefs and supply chain as the main restaurant, but with a different concept and branding. Thus, by matching data with operational knowledge, Deliveroo was able to co-create a successful brand that appeals to a different consumer base.
Another decision that heavily relies on data is selecting new markets to enter. Explaining the decision making process, Siddharth shared how consumers are always the first consideration for Deliveroo: “If we go there, what is the consumer need we will be solving, and will we do a good enough job?”
“At the end of the day, market share, profit, and so on, come from consumer need and if you’re the best placed to solve that need. There are many situations where we don’t feel we’re the best placed: we don’t have the right product, or the level of investment needed for the market. It might be a big market, but if we can’t build a product that is required for the consumer, we will never get in, our goal is to be number one or number two in every market that we’re in.”
The Asia Pacific performance hub
In 2019, Deliveroo set up a new Asia-Pacific performance hub in Singapore, its regional headquarters. Siddharth describes this as a work in progress, where they are evaluating ways to build their data science team out further, asking:
“What is the nature of the team that we need to build, how does it provide the right level of support, especially when they’re sitting further away from the engineers and so on.”
While each local team is data proficient, the new hub brings additional data skillsets and analytical firepower which Deliveroo looks to build for the APAC markets. With such large amounts of data generated and analysed to influence so many decisions, not processing it centrally in London is, at one level, a practical decision. But there are also differences in local markets that may not occur to a London based analyst. Weather patterns are one of them. As Siddharth shared:
The Singapore rain is unique: it doesn’t rain like that in London. It rains a lot in London but it just doesn’t rain like this. When it rains, like last Friday, no delivery is going to happen. With someone sitting in London looking at the metrics, they’ll ask ‘what happened here? how could things have gone so wrong?’ It’s just very hard for them to understand.”
The terrain’s influence on delivery times is also not always incorporated well into technology. For example, distances in Hong Kong may look short on a map, but it can actually contain many uphill and downhill paths or elevator trips, which affects the ability for the app to predict delivery times.
“Maps, including our internal mapping, is terrible at predicting the time on that or used to be. That needed a totally different technological intervention to try and understand.”
These may seem like small things but they matter to consumers. “Food”, says Siddharth, “is very emotional, so that leads to poor outcomes.” While the core product works, the company constantly thinks of local nuances and changes that they have to make which might not be applicable to the UK.
Some local factors that Deliveroo is looking at factoring in include how to add more local food and daily eating options. Siddharth shared that the considerations his company faces are how to do it in a sustainable way, that gives consumers better selection but also ensures that their flexibility and earnings are well taken care of. We can expect the same deep analysis of data and use of technology that has driven their previous ventures.