The Exchanges

Q in the microwave deal. So I'm totally with you on this one. But I do want to ask, you know, given it being your only business idea, but also the vision that you had for dark kitchens, you know, you started super early now, probably one of the hottest topics. So I guess where did that vision and strategic insight come from around dark kitchens and their importance moving forward?

A Look, the idea there, I don't think I had the idea day one when I had the business, but really I was kind of thinking about a few things, which was like, you build marketplace up by driving up supply and demand, right? Flywheel starts spinning. There are just some places where the supply wasn't great. And I was thinking like, ultimately, you know, how do you get that demand generation if you don't really have the supply? And so one idea was build delivery only kitchens, invite restaurants that maybe are popular in central London out to the suburbs. And effectively, you need supply in your marketplace then. Then on the operational side, it's much better because everything is geared towards delivery. So handoff time to the rider will be faster. The pick-pack process will be faster. And then people, ultimately, the restaurants are thinking delivery first. And so they're not trying to manage a dining room and kind of deal with delivery. They're thinking delivery first. And that yields all types of unanticipated innovations, like better packaging. And then I think, you know, for the restaurant itself, the P&L is really different. You don't have a front of house staff. Your rent is a fraction of what you would pay on the high street. It's all of these things, you know, we thought about sort of five years ago, and we've been incrementally building it. Today, some markets, it's a reall…

AI assessment note: “how do you get that demand generation if you don't really have the supply?”

Q But my question to you is like, what is really good in terms of driver drops per hour? How long does it take to get there? And what have been some lessons around kind of driver optimization on that standpoint?

A So we don't actually track rider drops per hour. And the reason we don't track it is, and I know I said it earlier, but so I'll backtrack and tell you initially, that's what we track. We don't track drops per hour because it's a number that's highly prone to gaming by the teams. Let me explain why. So what you really care about is drops per hour for a given radius and for a given sort of customer experience. And let me go into a little more detail what that means. I could say to you, well, Harry, you are now running the city of Leeds. I need you to get to 2.5 to three orders per hour, let's say. And you're like, fine. So if I give you the variable of, okay, you can start changing the delivery radii, you can start stacking orders. So you can say, you know what, instead of five kilometers, I'm going to make it two and a half, and I'm going to start picking two orders instead of one from Nando's. You can quickly get that to that number, but you're going to quickly compromise on the consumer experience, right? So that's one part. The second part is, you know, over 50% of our riders work on multiple platforms. And so what you actually care about is for any given one order, one order, single pickup, single drop off, what is the time the rider spends on that order? That's the metric we call rider experience time, and that's the number we're always trying to drive down.

AI assessment note: “So we don't actually track rider drops per hour. And the reason we don't track”

Q I mean, that is actually right. I know, it's scary how long I've been doing this, but I do want to start today with actually a question from Luciana, and I spoke to her before the show, and she asked, was Deliveroo your First business idea. And then secondarily, how did you come up with the idea for delivery given the hedge fund background?

A Yeah. So it is to this day, probably the only business idea I've had. I didn't set out to start a business per se. I set out to solve a consumer problem, started my career in wall street. No one in New York had a job where he worked a hundred hours a week, ordered every single meal, lunch, dinner, you know, at the office. And I was then transferred by Morgan Stanley, my employer to London. So this is Oh four. First day I get to work. Similar hours. I'm like, guys, what are we doing for dinner? Everyone's like, we're going to the Tesco in the Canary Wharf mall. We're getting microwave meals. And I was just like, no, no, no. This job is such a terrible job. Like the only fun part about it is your 20 pound allowance at dinner time. And so it was like kind of crushing to me. And it sounds pretty ridiculous that I'd care that much about food, but I do. And so literally day one in London is when I thought about the idea. And then for a while I was working at different jobs. I worked I worked at Morgan Stanley. I worked at a hedge fund. And so, and I actually tried starting the business in oh eight. When I thought about the whole process, I said, oh, I'd have to put a laptop in every restaurant. I'd have to invent some sort of Palm pilot device for our rider fleet. Obviously tablets and smartphones kind of were just coming into existence. And then in 2010, I went to Philadelphia for t…

AI assessment note: “So it is to this day, probably the only business idea I've had.”

Q But my question to you is like, what is really good in terms of driver drops per hour? How long does it take to get there? And what have been some lessons around kind of driver optimization on that standpoint?

A So we don't actually track rider drops per hour. And the reason we don't track it is, and I know I said it earlier, but so I'll backtrack and tell you initially, that's what we track. We don't track drops per hour because it's a number that's highly prone to gaming by the teams. Let me explain why. So what you really care about is drops per hour for a given radius and for a given sort of customer experience. And let me go into a little more detail what that means. I could say to you, well, Harry, you are now running the city of Leeds. I need you to get to 2.5 to three orders per hour, let's say. And you're like, fine. So if I give you the variable of, okay, you can start changing the delivery radii, you can start stacking orders. So you can say, you know what, instead of five kilometers, I'm going to make it two and a half, and I'm going to start picking two orders instead of one from Nando's. You can quickly get that to that number, but you're going to quickly compromise on the consumer experience, right? So that's one part. The second part is, you know, over 50% of our riders work on multiple platforms. And so what you actually care about is for any given one order, one order, single pickup, single drop off, what is the time the rider spends on that order? That's the metric we call rider experience time, and that's the number we're always trying to drive down.

AI assessment note: “So we don't actually track rider drops per hour. And the reason we don't track”

Q Can I ask, when you go into France, how do you decide where to start? How do you do zone selection? How do you do customer acquisition, driver acquisition? What does it take to launch a new market?

A So I think there are a few characteristics you look for first. And by the way, like the world's changed so much in six years, right? Or five, six years. But one thing we look at is population density. Population density, restaurant density can give you an indication of how many orders per hour you can do and ultimately give you a sense of how profitable a market can be from a gross profit standpoint. Second, we look at propensity to spend. Do people actually care about it? It's a funny thing, but not. Some markets are very different than others. For example, we found in Germany, what we found was that German people in many cases, like for dinner, they'll eat a black bread and cheese. But yeah, there are certain cultural aspects of certain places where maybe they just don't care as much. Another thing is the topography of the streets. Now, you want population density, but you don't want population density that's too high. Hong Kong, where we have a big business, you know, people go up these seventy-story buildings. Like, that's actually pretty challenging. Now, that's offset by the fact that you have so many restaurants in such a small area. What else do we look at? We look at competition. Who are the existing competitors? How well capitalized are they? Are they marketplace players or logistics players, right? And then we look at the labor environment. Are they going to allow fl…

AI assessment note: “one thing we look at is population density. Population density, restaurant density”

Q I mean, that is actually right. I know, it's scary how long I've been doing this, but I do want to start today with actually a question from Luciana, and I spoke to her before the show, and she asked, was Deliveroo your First business idea. And then secondarily, how did you come up with the idea for delivery given the hedge fund background?

A Yeah. So it is to this day, probably the only business idea I've had. I didn't set out to start a business per se. I set out to solve a consumer problem, started my career in wall street. No one in New York had a job where he worked a hundred hours a week, ordered every single meal, lunch, dinner, you know, at the office. And I was then transferred by Morgan Stanley, my employer to London. So this is Oh four. First day I get to work. Similar hours. I'm like, guys, what are we doing for dinner? Everyone's like, we're going to the Tesco in the Canary Wharf mall. We're getting microwave meals. And I was just like, no, no, no. This job is such a terrible job. Like the only fun part about it is your 20 pound allowance at dinner time. And so it was like kind of crushing to me. And it sounds pretty ridiculous that I'd care that much about food, but I do. And so literally day one in London is when I thought about the idea. And then for a while I was working at different jobs. I worked I worked at Morgan Stanley. I worked at a hedge fund. And so, and I actually tried starting the business in oh eight. When I thought about the whole process, I said, oh, I'd have to put a laptop in every restaurant. I'd have to invent some sort of Palm pilot device for our rider fleet. Obviously tablets and smartphones kind of were just coming into existence. And then in 2010, I went to Philadelphia for t…

AI assessment note: “So it is to this day, probably the only business idea I've had.”

Q Can I ask, when you go into France, how do you decide where to start? How do you do zone selection? How do you do customer acquisition, driver acquisition? What does it take to launch a new market?

A So I think there are a few characteristics you look for first. And by the way, like the world's changed so much in six years, right? Or five, six years. But one thing we look at is population density. Population density, restaurant density can give you an indication of how many orders per hour you can do and ultimately give you a sense of how profitable a market can be from a gross profit standpoint. Second, we look at propensity to spend. Do people actually care about it? It's a funny thing, but not. Some markets are very different than others. For example, we found in Germany, what we found was that German people in many cases, like for dinner, they'll eat a black bread and cheese. But yeah, there are certain cultural aspects of certain places where maybe they just don't care as much. Another thing is the topography of the streets. Now, you want population density, but you don't want population density that's too high. Hong Kong, where we have a big business, you know, people go up these seventy-story buildings. Like, that's actually pretty challenging. Now, that's offset by the fact that you have so many restaurants in such a small area. What else do we look at? We look at competition. Who are the existing competitors? How well capitalized are they? Are they marketplace players or logistics players, right? And then we look at the labor environment. Are they going to allow fl…

AI assessment note: “one thing we look at is population density”

Q Question for you is like, you as a leader have to change with every element of scale. How have you changed as a leader now you're a public market CEO?

A Yeah, look, I think that there are a few things that have changed and a few things that are probably the same. I think I'm a straightforward person who is transparent, and that's an important quality that I want out of my team. I always tell people what's going on, whether it's good, whether it's bad, whether it's neutral, but just do that. I think in In terms of the way I've changed, I certainly, you know, I will process a bit before I answer questions. Obviously, we're public, so you can't just say whatever comes, you know, in your mind, so you sort of think about it a bit. You sort of think about the implications of what you say, especially when you're a leader of 2500 people, and obviously we work with, you know, hundreds of thousands of writers every day as well. The other thing I do very deliberately now is put myself in the shoes of a junior level, mid-level employee, and say, if Will says something, How are they going to interpret it? And ultimately, is that going to help them do their job better? And that's something that comes with experience. That's not something that is going to come from first principles problem solving when there's like three people at the company. You kind of just become conscious of it over time and realize your voice is pretty powerful. You need to use it the right way.

AI assessment note: “In terms of the way I've changed, I certainly... process a bit before I answer”

Q We mentioned France, we mentioned new markets. In terms of existing markets, and I wasn't sure whether I could ask this, but I was told I could, so this is great. You know, Uber Eats came and they came hard. Tell me, how was that experience for you when it happened? What were some learnings when they were launching and when you were fighting against them?

A Yeah, look, I think that this business is an interesting one because it is A hyper-local business, yet you're fighting against, you know, global internet giants, right? So it's this kind of weird dichotomy, and it's also a physical business. And I think that, you know, Uber, you know, has been a great competitor, you know, ever since they got into this business, you know, we compete in various markets. I think what they came in with back, I think it was, I don't fully remember. You know, they came in with a secret weapon around the world, right? And I'm sure Tony from DoorDash and all these other guys will tell you the same thing. They had McDonald's. Right. They had this, uh, secret weapon and McDonald's is the world's most famous food brand. So they came in, you know, they gave a lot of stuff away for free, you know, kind of all this textbook playbook. And what I realized was we were like, oh my God, like this is going to be really hard. And it was right. No question. You had this like pretty hardcore competitor come in, but then we realized something. We realized, you know what, at the end of the day, there's no shortcuts in this business. It's all about your hyper local consumer value proposition. Do you have the right restaurants? Are you doing deliveries? You know, in the time you said you would, are you communicating all of this stuff clearly to the customer through the …

AI assessment note: “we realized, you know what, at the end of the day, there's no shortcuts”

Q What were the signals that it was working well to you?

A All right, a few. So, I'll give you a few, one funny one and one sort of statistical one. So, Initially, we were in one neighborhood, right? South Kent, Chelsea. And I used to beg my friends to order food. That was my demand gen channel. And they'd be like, yeah, fine. And then I realized they would order food only because they thought it was funny that I would deliver the food. And then after a while, I just noticed they kept ordering. And I'm like, oh, so maybe they actually just like the service, right? So that was one funny one. The more statistical one was once we got to zone three in London, of course, zone three is very affluent, but it's not zone one. The retention and frequency of the customers was exactly the same. And I was like, oh, okay. If it works here, then I think it's going to work in a lot of places. Then we saw that in Bright. We saw it in Cambridge. And I was like, oh, well, this is not London. I mean, there are affluent places, but they're different. And then we were like, you know what, we're going to go for this internationally.

AI assessment note: “I'll give you a few, one funny one and one sort of statistical one”

Q that? Because we're seeing a lot of, you know, quick commerce players now expand so fast across so many cities. And I think from what I heard from your ambassadors, you were very deliberate and strategic around acquiring supply And moving from zone to zone. How did you think about that zone to zone expansion? And how were you kind of strategic and deliberate around it in the early days?

A It's funny. I don't know that it was deliberate or strategic. I funded the company myself. It's not like we had any money. It was a very different fundraising environment in 2012. So I had savings, you know, I put in 30,000 dollars or pounds. Um, and I'm like, okay, this better last for a year on if this doesn't really work, I gotta like figure something out. And it was really a Capital issue as opposed to anything else. Cause once we figured out how everything worked, it was like, okay, we got to expand. We got to raise some money. Right. So it wasn't delivered at all. I mean, I did the most unscalable things. I did deliveries for an entire year. I, you know, signed up every restaurant at nighttime worked on product. Right. So it was all like, it just was what it was. Right.

AI assessment note: “I don't know that it was deliberate or strategic. I funded the company myself.”

Q What were some of your big lessons around zone maturity and speed to zone break even, given the expansion that you saw?

A What we realized pretty quickly was driving density of orders Um, would actually make the unit economics work pretty quickly, but it's kind of hard to do that all at once, you know, let's say 200 cities, because this is a hyper local marketplace, right? So you, as the customer, Harry, you know, living in London, you order from us due to the strength of our local offering and the strength of our local logistics, right? You don't care what we do in East London. You don't care what we do in Paris. You care about your neighborhood. And getting that right is a bit of a time consuming process. And then doing that kind of across many, many markets at once is hard, right? It requires a lot of initially just feet on the ground and figuring it out. And then eventually as you build these playbooks and you become much more standardized, you can automate a lot of those things. That definitely wasn't the case initially, right? You're just trying to figure it out.

AI assessment note: “driving density of orders Um, would actually make the unit economics work pretty quickly”

Q Question for you is like, you as a leader have to change with every element of scale. How have you changed as a leader now you're a public market CEO?

A Yeah, look, I think that there are a few things that have changed and a few things that are probably the same. I think I'm a straightforward person who is transparent, and that's an important quality that I want out of my team. I always tell people what's going on, whether it's good, whether it's bad, whether it's neutral, but just do that. I think in In terms of the way I've changed, I certainly, you know, I will process a bit before I answer questions. Obviously, we're public, so you can't just say whatever comes, you know, in your mind, so you sort of think about it a bit. You sort of think about the implications of what you say, especially when you're a leader of 2500 people, and obviously we work with, you know, hundreds of thousands of writers every day as well. The other thing I do very deliberately now is put myself in the shoes of a junior level, mid-level employee, and say, if Will says something, How are they going to interpret it? And ultimately, is that going to help them do their job better? And that's something that comes with experience. That's not something that is going to come from first principles problem solving when there's like three people at the company. You kind of just become conscious of it over time and realize your voice is pretty powerful. You need to use it the right way.

AI assessment note: “In terms of the way I've changed, I certainly, you know, I will process a bit”

Q that? Because we're seeing a lot of, you know, quick commerce players now expand so fast across so many cities. And I think from what I heard from your ambassadors, you were very deliberate and strategic around acquiring supply And moving from zone to zone. How did you think about that zone to zone expansion? And how were you kind of strategic and deliberate around it in the early days?

A It's funny. I don't know that it was deliberate or strategic. I funded the company myself. It's not like we had any money. It was a very different fundraising environment in 2012. So I had savings, you know, I put in 30,000 dollars or pounds. Um, and I'm like, okay, this better last for a year on if this doesn't really work, I gotta like figure something out. And it was really a Capital issue as opposed to anything else. Cause once we figured out how everything worked, it was like, okay, we got to expand. We got to raise some money. Right. So it wasn't delivered at all. I mean, I did the most unscalable things. I did deliveries for an entire year. I, you know, signed up every restaurant at nighttime worked on product. Right. So it was all like, it just was what it was. Right.

AI assessment note: “I don't know that it was deliberate or strategic. I funded the company myself.”

Q I presume it's because it's, you know, under 30 minutes, you know, unless you're using, you know, a terrible provider. But, like, I think that would be the reasoning. My question to you on this is, like, quite simply, how do you analyze the rise of this space? Is it competitive to deliver? How do you think about this?

A I think that if you think about the tan of the 12 markets we're in between grocery and restaurant food, it's 1.2 trillion pounds or 1.8 trillion dollars, right? So it's a really big space. I think there are many different use cases for restaurant food or grocery. So for example, if you're celebrating your birthday, you're going to go to a restaurant. If you're tired after the week and you know, you're home with a few friends, you might order delivery. If you're a family and you want to order groceries for a large group of people and you have a very regular schedule, Ocado is great. Tesla is a great product for that. You know, but if you want something pretty quick, you know, our grocery business, I think we said in our H-one results is over seven percent of our business now, right? It's a big part of our business, which is our model is a little bit different than the dark store guys, but point is, you know, I think people are expecting things faster, and they're also kind of approaching it from different occasions. So if I want to cook a pasta dinner for me and my friends, you can actually get that in 25 minutes now, right? The ingredients. You don't have to order two days in advance like you did in the past, right? Or go to the store. So I think that The short answer is all of these different players, us, Okado, Getir, whoever else, Gorillaz, all of this stuff, we all play a p…

AI assessment note: “based on the occasion, you know, we may compete, but we also may not compete.”

Q No, I totally get you. I do want to finish on the next five years for you and Deliveroo. What does that look like? And how big could this be, Will?

A You know, as I said, we've been adding new verticals in every year, right? So whether that's starting with independent restaurants, then we added casual dining chains, then we added fast food, then we added grocery, we add additions, we add signature, we add plus, and we are still at this .3 to five percent penetrated, right? Online penetrated. So we are still super early in this journey. And this journey of online food is very difficult to solve. Because it's perishable, and it's also emotional, right? And getting that right online is just hard. So I think we're super early, and so that's the problem, you know, we'll be solving. And as for me, I love what I do. I want to keep doing this and just keep solving these consumer problems. And when I say consumer, it's the three sides of our marketplace. I'm just really passionate about it.

AI assessment note: “we are still at this .3 to five percent penetrated, right? Online penetrated.”

Q in the microwave deal. So I'm totally with you on this one. But I do want to ask, you know, given it being your only business idea, but also the vision that you had for dark kitchens, you know, you started super early now, probably one of the hottest topics. So I guess where did that vision and strategic insight come from around dark kitchens and their importance moving forward?

A Look, the idea there, I don't think I had the idea day one when I had the business, but really I was kind of thinking about a few things, which was like, you build marketplace up by driving up supply and demand, right? Flywheel starts spinning. There are just some places where the supply wasn't great. And I was thinking like, ultimately, you know, how do you get that demand generation if you don't really have the supply? And so one idea was build delivery only kitchens, invite restaurants that maybe are popular in central London out to the suburbs. And effectively, you need supply in your marketplace then. Then on the operational side, it's much better because everything is geared towards delivery. So handoff time to the rider will be faster. The pick-pack process will be faster. And then people, ultimately, the restaurants are thinking delivery first. And so they're not trying to manage a dining room and kind of deal with delivery. They're thinking delivery first. And that yields all types of unanticipated innovations, like better packaging. And then I think, you know, for the restaurant itself, the P&L is really different. You don't have a front of house staff. Your rent is a fraction of what you would pay on the high street. It's all of these things, you know, we thought about sort of five years ago, and we've been incrementally building it. Today, some markets, it's a reall…

AI assessment note: “I was kind of thinking about a few things, which was like, you build marketplace”

Q Can I ask, you know, if we look at international markets in particular, I heard the French market was an interesting one because You were a bit of a later entrant into the French market. How did you win it, and what strategic decisions do you think you made to allow yourself to beat existing players in a market that wasn't your home market?

A Well, I don't know if we've won any market to date. I mean, this is such a dynamic industry, and I don't know what winning means exactly either. The reason it's important is online penetration of food is three to five percent still today. Now, I think the US sort of leapfrogged a lot during COVID, but if you look at Europe, It's still in that, you know, one meal out of 21 a week is online. So I think we're, we're super, super early. But let me give you the example of France, which I think is pretty interesting. So we went into France in maybe late 20 15, early 16. There were a few players in the market that were primarily marketplace players. And what I mean by that is they relied on the restaurant for filling the order, like the old Just Eat model here. And French people, you know, they're pretty demanding when it comes to food. And when you think of France, What's interesting, and this is me, like, 20 years ago, there were Michelin-starred restaurants, brasseries, Vietnamese restaurants, and kebabs, right? That was kind of like, at least back then, 20 years ago. But when we went into the market, we noticed something changed completely in Paris. The rise of casual dining made a big difference. So in Paris, five, six years ago, you had the Parisian take on a taco, the Parisian take on a burger. These are all-day menus, limited in scope, very focused, Very casual. Having that as…

AI assessment note: “There were a few players in the market that were primarily marketplace players.”

Q I'm interested with that in mind that it's like a weird negative network effects business, where like the more successful you become in terms of consumer demand, the more likely That a driver is to take two or three orders, say, and then it reduces the consumer experience. How do you think about that?

A You just can't do it. You can't do it. Like, here's the thing. Like, you asked about our delivery kitchens, right? That rider experience time is significantly lower because everything's geared towards delivery. So it's the handoff, the pick pack, all that stuff. How quickly the guy waits around is much lower. That's, that's like one way. The other way is actually the machine learning algorithms when it comes to something that we call prep time for restaurants. I'll give you an example. So like Nando's, let's say around the office, what is the time is it now? Let's say around six PM on a Monday evening, Nando's near the office. We have so many data points that we know for a 25 pound order, we have a very good sense of how long it's going to take to cook. And then we send the rider at the right time to minimize his or her wait time there. But Nando's is a very popular restaurant. What if, you know, Harry's pizza place in Marseille, we may not have that many data points, right? And so the variability there is going to be a bit higher. And as we get more data points, you know, our machine learning models will get better and better, and then ultimately we can try to predict that prep time to minimize your wait time at the restaurant.

AI assessment note: “You just can't do it. You can't do it.”

Q How important do you think exclusivity on the supply side was? You know, I love nowadays, for example, so I would only ever go to delivery. How important was the exclusive to, you know, really standing up and being a match competitor?

A I believe an online food platform is an emotional platform. I think food is emotional, and I think that, you know, having the right brands on your platform, highlighting them, and if you can have an advantage over someone else because, you You have a brand that they don't, and you've developed a relationship with a restaurant over a very long period of time, and they have that trust in you to, to deliver that. Yes. I think that's, you know, all things being equal, you know, a good thing. And that's why we build things like additions kitchens that are for our exclusive partners, right? So the way we sort of think about it is how can we drive maximum value for our partners in a way that expands their business in a way that makes it more profitable for them. So we're really thinking about how can we Tie ourselves in to be a strategic partner as opposed to a transactional partner. And what we've seen is, you know, when that goes really, really well, people are like, oh, you know, it would be nice to work with you on an exclusive basis.

AI assessment note: “if you can have an advantage over someone else because... You have a brand that they don't”

Q from two points. One is when I spoke to Fred beforehand, he mentioned you were raising, I think it was about two hundred and fifty million dollars from the bridge points and the general catalyst of the world. And you did it without a banker, which is pretty unheard of, given the size of the raise. Why did you decide to do that, Will? And how did it play out?

A I'm just thinking, so that was, that must have been the series, man, I don't remember anymore, maybe. That raise was pretty tricky, actually, because A, Brexit had just happened. So people were like, I remember this well, Brexit had just happened, and then Uber Eats said to every country in the world. So those two things happened at the same time. And I remember the Bridgepoint had, I don't know if they had signed a term sheet or they were close to it, but I just remember, you know, General Catalyst and all these guys, and people are like, whoa, like, what's going on here? So that was pretty hard, but I think at the end of the day, I just went in there and said, does this really change kind of your perspective on what we're trying to do? We're really, really early, and then, you know, what does really Brexit mean when our business is a hyper-local business? We're not a sort of import-export business, and so there's a lot of conversations there, but it was tough. Look, we didn't use a banker because we never used a banker. I come from finance, so I understand, you know, the concept of a participating preferred, and, you know, I can have a sort of high-level understanding of this stuff. I'm not like a pro, but I think, look, at the end of the day, what is a banker good for? I think bankers are great for running processes. They're great for, you know, driving sort of competitive t…

AI assessment note: “we didn't use a banker because we never used a banker. I come from finance”

Q I presume it's because it's, you know, under 30 minutes, you know, unless you're using, you know, a terrible provider. But, like, I think that would be the reasoning. My question to you on this is, like, quite simply, how do you analyze the rise of this space? Is it competitive to deliver? How do you think about this?

A I think that if you think about the tan of the 12 markets we're in between grocery and restaurant food, it's 1.2 trillion pounds or 1.8 trillion dollars, right? So it's a really big space. I think there are many different use cases for restaurant food or grocery. So for example, if you're celebrating your birthday, you're going to go to a restaurant. If you're tired after the week and you know, you're home with a few friends, you might order delivery. If you're a family and you want to order groceries for a large group of people and you have a very regular schedule, Ocado is great. Tesla is a great product for that. You know, but if you want something pretty quick, you know, our grocery business, I think we said in our H-one results is over seven percent of our business now, right? It's a big part of our business, which is our model is a little bit different than the dark store guys, but point is, you know, I think people are expecting things faster, and they're also kind of approaching it from different occasions. So if I want to cook a pasta dinner for me and my friends, you can actually get that in 25 minutes now, right? The ingredients. You don't have to order two days in advance like you did in the past, right? Or go to the store. So I think that The short answer is all of these different players, us, Okado, Getir, whoever else, Gorillaz, all of this stuff, we all play a p…

AI assessment note: “based on the occasion, you know, we may compete, but we also may not compete.”

Q from two points. One is when I spoke to Fred beforehand, he mentioned you were raising, I think it was about two hundred and fifty million dollars from the bridge points and the general catalyst of the world. And you did it without a banker, which is pretty unheard of, given the size of the raise. Why did you decide to do that, Will? And how did it play out?

A I'm just thinking, so that was, that must have been the series, man, I don't remember anymore, maybe. That raise was pretty tricky, actually, because A, Brexit had just happened. So people were like, I remember this well, Brexit had just happened, and then Uber Eats said to every country in the world. So those two things happened at the same time. And I remember the Bridgepoint had, I don't know if they had signed a term sheet or they were close to it, but I just remember, you know, General Catalyst and all these guys, and people are like, whoa, like, what's going on here? So that was pretty hard, but I think at the end of the day, I just went in there and said, does this really change kind of your perspective on what we're trying to do? We're really, really early, and then, you know, what does really Brexit mean when our business is a hyper-local business? We're not a sort of import-export business, and so there's a lot of conversations there, but it was tough. Look, we didn't use a banker because we never used a banker. I come from finance, so I understand, you know, the concept of a participating preferred, and, you know, I can have a sort of high-level understanding of this stuff. I'm not like a pro, but I think, look, at the end of the day, what is a banker good for? I think bankers are great for running processes. They're great for, you know, driving sort of competitive t…

AI assessment note: “we didn't use a banker because we never used a banker. I come from finance”

Q the topography there. I'm interested. I think people don't Often understand fully the complexity from a technical standpoint with which you have to address driver distribution, driver optimization from pickup to delivery. But from a technical standpoint, what was some of the big hurdles and what was kind of the machinery that you had to build behind the scenes to handle? Very different topography, very different landscapes. What was that?

A I would say at first, you know, we were super simplistic. You know, our algorithm is what you would call initially was a greedy solver, which basically meant it would assign the closest available rider. To an order. Now over time, we obviously, we implemented machine learning models pretty soon after. The point is, you know, the models will retrain themselves. There's a lot more data feeding them. And so while a greedy solver can get you maybe initially 70% of the way there, that's fine for a while. Layering on, you know, all of this, you know, additional machinery made it, you know, much more profitable over time. I think, you know, and embedded in those models, you know, a lot of things that you're asking about, right? So we're running a lot of different models at the same time.

AI assessment note: “our algorithm is what you would call initially was a greedy solver”

Q Can I ask, you know, if we look at international markets in particular, I heard the French market was an interesting one because You were a bit of a later entrant into the French market. How did you win it, and what strategic decisions do you think you made to allow yourself to beat existing players in a market that wasn't your home market?

A Well, I don't know if we've won any market to date. I mean, this is such a dynamic industry, and I don't know what winning means exactly either. The reason it's important is online penetration of food is three to five percent still today. Now, I think the US sort of leapfrogged a lot during COVID, but if you look at Europe, It's still in that, you know, one meal out of 21 a week is online. So I think we're, we're super, super early. But let me give you the example of France, which I think is pretty interesting. So we went into France in maybe late 20 15, early 16. There were a few players in the market that were primarily marketplace players. And what I mean by that is they relied on the restaurant for filling the order, like the old Just Eat model here. And French people, you know, they're pretty demanding when it comes to food. And when you think of France, What's interesting, and this is me, like, 20 years ago, there were Michelin-starred restaurants, brasseries, Vietnamese restaurants, and kebabs, right? That was kind of like, at least back then, 20 years ago. But when we went into the market, we noticed something changed completely in Paris. The rise of casual dining made a big difference. So in Paris, five, six years ago, you had the Parisian take on a taco, the Parisian take on a burger. These are all-day menus, limited in scope, very focused, Very casual. Having that as…

AI assessment note: “Having that as a prerequisite really allowed the delivery space to take off.”

Q So without experience, would you say to founders now do the unscalable things to scale? And I'm sure you learned a ton from doing that yourself. Or would you actually say with capital availability as it is run?

A Man, I don't know. I don't want to sound like some old guy, you know, like, oh, there's so much capital out there, and things are different, but it is different, right? Like, I'll tell you a funny story. So Greg, this is actually kind of an insane story. It might actually bore you, but I'll tell it to you anyway. So we initially, you know, we didn't know, like, what tablets to use, but we were super cheap, right? So we're like, we're not buying iPads. I get on eBay or Amazon. I'm looking for the Asus seven. I think that's what it was, an Asus seven or eight. Use tablets, you know, 40 pounds, right? So we'd buy these, we'd repurpose them, you know, completely flash the tablet, do everything manually. One by one, we would flash these tablets and, and put our restaurant order manager on it. And then Greg was like, you know what? We need a stand, right? Everything needs a stand in the restaurant, right? And instead of like us buying these, he decided to, I still cannot believe he did this. He bought these microphone stands, like used, like in huge quantities, but you couldn't just attach it with duct tape. So what he actually did Dended is he forced me to assemble a three D printer. Okay. And we would print these parts out that would fit a fix the tablet onto this mic stand. The printer would print out one part a day. And I had to get up at three in the morning to reschool the stup…

AI assessment note: “there was a lot of stuff that wasn't necessary, but I will say this”

Q the topography there. I'm interested. I think people don't Often understand fully the complexity from a technical standpoint with which you have to address driver distribution, driver optimization from pickup to delivery. But from a technical standpoint, what was some of the big hurdles and what was kind of the machinery that you had to build behind the scenes to handle? Very different topography, very different landscapes. What was that?

A I would say at first, you know, we were super simplistic. You know, our algorithm is what you would call initially was a greedy solver, which basically meant it would assign the closest available rider. To an order. Now over time, we obviously, we implemented machine learning models pretty soon after. The point is, you know, the models will retrain themselves. There's a lot more data feeding them. And so while a greedy solver can get you maybe initially 70% of the way there, that's fine for a while. Layering on, you know, all of this, you know, additional machinery made it, you know, much more profitable over time. I think, you know, and embedded in those models, you know, a lot of things that you're asking about, right? So we're running a lot of different models at the same time.

AI assessment note: “embedded in those models, you know, a lot of things that you're asking about”

Q So without experience, would you say to founders now do the unscalable things to scale? And I'm sure you learned a ton from doing that yourself. Or would you actually say with capital availability as it is run?

A Man, I don't know. I don't want to sound like some old guy, you know, like, oh, there's so much capital out there, and things are different, but it is different, right? Like, I'll tell you a funny story. So Greg, this is actually kind of an insane story. It might actually bore you, but I'll tell it to you anyway. So we initially, you know, we didn't know, like, what tablets to use, but we were super cheap, right? So we're like, we're not buying iPads. I get on eBay or Amazon. I'm looking for the Asus seven. I think that's what it was, an Asus seven or eight. Use tablets, you know, 40 pounds, right? So we'd buy these, we'd repurpose them, you know, completely flash the tablet, do everything manually. One by one, we would flash these tablets and, and put our restaurant order manager on it. And then Greg was like, you know what? We need a stand, right? Everything needs a stand in the restaurant, right? And instead of like us buying these, he decided to, I still cannot believe he did this. He bought these microphone stands, like used, like in huge quantities, but you couldn't just attach it with duct tape. So what he actually did Dended is he forced me to assemble a three D printer. Okay. And we would print these parts out that would fit a fix the tablet onto this mic stand. The printer would print out one part a day. And I had to get up at three in the morning to reschool the stup…

AI assessment note: “there was a lot of stuff that wasn't necessary, but I will say this”