DASH - NASDAQ

Thank you for standing by, and welcome to the DoorDash Q2 2025 Earnings Call. [Operator Instructions] I would now like to turn the call over to Wes Twigg. Please go ahead, Wes.

Good afternoon, everyone, and thanks for joining us for our Q2 '25 earnings call. I'm pleased to be joined today by Co-Founder, Chair and CEO, Tony Xu; and CFO, Ravi Inukonda. We'll be making forward-looking statements during today's call, including, without limitation, our expectations for our business, financial position, operating performance, profitability, our guidance, strategies, capital allocation approach and the broader economic environment. Forward-looking statements are subject to risks and uncertainties that could cause actual results to differ materially from those described. Many of these uncertainties are described in our SEC filings, including our most recent Form 10-K and 10-Qs. You should not rely on our forward-looking statements as predictions of future events or performance. We disclaim any obligation to update any forward-looking statements, except as required by law. During this call, we will discuss certain non-GAAP financial measures. Information regarding our non-GAAP financial measures, including a reconciliation of such non-GAAP measures to the most directly comparable GAAP financial measures may be found in our earnings release, which is available on our Investor Relations website at ir.doordash.com. These non-GAAP measures should be considered in addition to our GAAP results and are not intended to be a substitute for our GAAP results. Finally, this call is being audio webcasted on our Investor Relations website. An audio replay of this call will be available on our website shortly after the call ends. Operator, I'll pass it back to you, and we can take our first question.

[Operator Instructions] And your first question comes from the line of Shweta Khajuria with Wolfe Research.

Let me try 2, please. On the U.S. marketplace, orders accelerating year-over-year. You continue to see strength in restaurants. So it sounds like the strength was driven by DashPass membership growth, frequency uptick and potentially some product improvements around personalization and maybe faster delivery. So I guess my question is, could you perhaps point to some specific examples around what you did on personalization and what specifically drove strength in membership growth and frequency? And then the second question is on advertising revenue. You've now exceeded $1 billion in annualized revenue run rate, and you got there fairly quickly. So post-Symbiosys acquisition, could you please share your thoughts on how you're thinking about scaling the on- platform and off-site ads opportunity as you see it in the near to midterm?

Shweta, it's Tony. I'll start, and Ravi, feel free to chime in. On the first question about the accelerated growth at bigger scale, look, a lot of these improvements and developments were the results of our team's work, probably dating back a couple of years ago. I've always believed that it's very, very difficult to make changes, especially at our scale that can have material impact within a short time period, whether it's the quarter or even the year. And in many ways, that goes to the work that we do today, which is the work that our teams are working on right now likely will have an impact on a quarter a few years into the future. And I think if we're doing our work right, that is always going to be the cadence here at DoorDash, where the North Star focus is always to make improvements, first and foremost, to our products, to the selection that we offer of both stores and items, to the affordability of those items and stores, the quality of delivery, the timeliness, the accuracy and certainly, the customer service, especially when we get things wrong. And I think that's been a consistent narrative that reflects a consistent execution at the company for the 12 years that we've been doing this, and it remains to be the case today. So we've seen improvements across the board in our various products that have achieved the results that you're seeing today. With respect to the second question on ads, you're right, at some point last year, we did cross $1 billion of revenue run rate in the ads business, making it the fastest in history to get there. And I mentioned all along that the ads business has been progressing really healthily. But the focus for the ads business today and 3 years ago when we started the business and in the future will always remain the same, which is we have to achieve the best consumer experience and build the most successful marketplace, which is what will enable us to build the most successful return on ad spend product or the most successful advertising business for retailers, CPGs and restaurants. That remains the focus. With respect to Symbiosys, an acquisition that we closed, one of the developments we've always had is that we've always been privileged where there is more dollars that advertisers want to spend with us than we are willing to give in terms of the surfaces in which we allow them to advertise. Again, we're trying to make sure that we have a best-in-class consumer experience. And therefore, whether it's compared to peers or just compared to what we think the right thing to do, is we don't allow advertisements to just show up in the products. That doesn't mean though that there isn't a good use of dollars or proceeds from these advertisers. This is why we purchased Symbiosys, where together with our know-how as well as our information that we collect, we believe that we can be the best marketer. You can almost think of it as we are the agent of choice where we can do the best marketing spend for each one of these merchants and advertisers. A lot of times, I think when you think about our relationship with everyone we work with, we are not just a delivery provider. Many times, we are the growth engine or the growth button, both in terms of what we can bring incrementally from our marketplace as well as the know-how that we bring from our first-party business in teaching these retailers and advertisers on how to build the most successful digital powerhouses in their own rights. And so you see that again here in the Symbiosys example as part of our advertising business.

Shweta, it's Ravi. I think one of the things that you mentioned was the acceleration of the ads business. I mean I would clarify, right, I would not read into the disclosure as us changing the way in which we are operating the ads business. We're operating the business with the same level of discipline. The 2 metrics are merchant ROAS as well as consumer conversion. We think we are best-in-class for both of those. The goal for us has always been to build a great marketplace business. And if we do that, the ads business will scale and will continue to grow, but we are growing it with the same level of discipline as we've done before.

And your next question comes from the line of Deepak Mathivanan with Cantor Fitzgerald.

Tony, I wanted to ask an AI question. We're seeing big improvements in search and recommendation systems across a lot of consumer Internet platforms basically using AI models with reasoning capabilities. You obviously have a very strong engineering team that's on top of these tech breakthroughs. Can you talk about some of the ways Dash is planning to use these larger AI models to both improve user experience and also find additional pockets of operational efficiencies, maybe, say, in the next 6 to 12 months? And then the second question, maybe for Ravi. With all the supply growth and value prop improvements that you've done with new verticals, how meaningfully is new verticals now contributing to kind of new customer growth to the platform? And also, maybe talk a little bit about the cross-platform benefits for the U.S. restaurant business from new customers acquired through grocery and perhaps convenience.

Sure. Deepak, on the first question on AI, I'll start. Whenever you see a revolutionary technology that seems to be improving by the hour, you kind of have to rethink your entire business. And I think this is a thought that Don and me probably 4 years ago when you started seeing kind of the beginnings of what these things can do because I think that in many ways, it's not just a new technology, it's also a new way of how you attack different problems. So there are a few areas in which we found a lot of fruit in terms of how we thought about applying it. So the first is thinking about what should the consumer product actually look like in a world where, to your point, there's a different way of doing everything from search to personalization to ranking. And I would say that there's a lot of different iterations that I think you'll see us take in terms of shipping a version of the product that will really, I think, marry the best of what the technology brings in a way that actually brings pragmatic benefits to different use cases to the consumer. The second area is a lot of our business happens in the physical world. It's not just the deliveries, whether it's the creation of inventory lists or menus from restaurants or all the work that we do to onboard different merchants and retailers and advertisers, there is a lot of physical activity or repetitive activity that happens. All of those processes can be and are being rethought right now at DoorDash in terms of how we can do that in much more efficient ways with higher quality and fidelity. And the final way, and we see this most notably in our engineering teams, is really how do we become a more productive organization, if we were to start this company from scratch in 2025 versus getting our start back in 2013. We see this a lot in coding, certainly, where obviously, probably not a surprise, when you're talking about large language models, when you're talking about domains where there are correct answers and easy ways to produce evaluations that can tell you whether you're on the right track or wrong track, coding has been a natural use of productivity for DoorDash now for a few years in a row. And I think that's only going to increase. And that should also be true, though, in other functions. If I were building DoorDash from scratch today, we would have a very different approach given all the tools available. But this is kind of like when DoorDash started 12 years ago where you had the real acceleration of cloud computing services where DoorDash no longer needed to manage data centers. We can kind of get easy access to AWS instances and kind of get up and running. This is kind of true today, but just at the functional level instead of just the compute level, where across every discipline inside the company, I think you can, from first principles, re-derive how you would build the company from scratch, and that's what we're thinking about.

Deepak, it's Ravi. I'll talk about new verticals, right? I know your question was around new CX, but let me take a step back and give you how we're thinking about overall new verticals and the performance of the business. I mean, look, new verticals had a really strong quarter. They're growing really fast, growing much faster than our core restaurants business. In fact, we grew share. And I think like we said last time, our expectation is that we'll be volume share leaders within the next year. If you actually break that apart, we are seeing strength both from new as well as existing cohorts. On the new cohorts, today, the size of new cohorts ordering from new verticals is higher than same time last year. The engagement levels of mature cohorts, think of them as existing cohorts, the size of those cohorts is also larger than what we saw last year. At the same time, these cohorts are engaging with us more, which means order frequency is continuing to increase. So when you look at the underlying cohort performance, it's very strong. Both users are growing, order frequency is growing. We've hit all-time highs across both of those metrics. The other question is what's actually driving that growth, right? If you look at the performance of the business, we've increased the selection on the platform quite considerably. The quality of the product is continuing to get better. We've made the product more affordable. More DashPass users are ordering from more categories. Look, in Q4, we talked about the fact that roughly about 1/4 of our users order from new verticals. That number has continued to increase. All of these are signs pointing to cohorts getting value, consumers getting value from the product. And for us, the focus continues to be to improve the underlying value proposition. And as long as we continue to do that, we are confident that it's going to be a large business for us over time.

And your next question comes from the line of Ron Josey with Citi.

I wanted to ask a little bit more. The letter or the press release talks quite a bit about cohort trends, particularly within DashPass members as you still continue to have records. So talk to us a little bit more about how DashPass cohorts have been trending here, maybe on the newer cohorts in the last year or 2, are they witnessing similar results? And then as we think about sort of just the maybe bigger picture, and I know we're still yet to close Deliveroo, but any updated thoughts on the integration there?

Yes, Ron, I'll take the first one on DashPass and Tony, feel free to chime in. Look, I mean, DashPass, Ron, had a very, very solid quarter. Look, I mean, we've been investing in DashPass for years, right? Ultimately, the thing that is driving growth for DashPass for us is the underlying product continuing to get better. We're focused on adding more selection. We're focused on driving higher quality. We're making the product continue to be more affordable. There's more features and services available as a part of DashPass. Ultimately, all of this is driving the user journey. What we see is we're acquiring more new consumers than ever before. Those consumers are joining the platform. They're habituating more. They continue to graduate to DashPass. And DashPass, when they continue to graduate, the order frequency is higher. One of the things that we're seeing from a cohort perspective, Ron, in DashPass, even the older cohorts, even as old as 5, 6 years ago, those cohorts continue to engage with us at higher levels. All of this is pointing to the underlying strength that we are seeing in this business. I mean, look, it's not anything that we've done over the last couple of quarters, right? We've been investing in the product for several years now. The underlying product is getting better. That's what's driving the strength both in DashPass as well as the underlying cohorts.

And Ron, it's Tony. I think your second question is about Deliveroo. Probably, obviously, there's not much I can say about it besides the fact that we anticipate closing on the time line we had communicated, which is at some point in Q4, it's obviously still subject to regulatory review. But look, with any acquisition, I mean, we're going to run the company, should it close in the way that we run anything at DoorDash, which is, first and foremost, we have to build and invest in the best product experience. And if we can do that and improve the order rates as well as improve the unit economics where we see efficiency gaps, those are the tools in which we can use to reinvest back into each one of the audiences. Again, our investment philosophy has always been one of maximizing long-term margin dollars, not short-term unit margins. And that will continue to be the focus that we take to the Deliveroo project, should it close, and any other project at DoorDash.

And your next question comes from the line of Nikhil Devnani with Bernstein.

Tony, are you surprised by the ongoing strength in the size of new customer cohorts in the U.S. I mean it would seem like your audience in the U.S. is pretty large already, in the high tens of millions of people, which is a big number relative to U.S. households, but that doesn't seem to be affecting the funnel for new customer acquisition at all. So when you step back and take a multiyear view, I mean, how durable do you think this trend is? And are you able to also break down how much growth this business gets from new customers versus pre-existing cohorts in a given year?

Yes. Nikhil, no, I'm not surprised is the short answer. And I think I may have said this in a few different calls before, whether when we were discussing this during COVID, peak inflation in '21, the normalization of COVID or COVID reopenings in '22, where -- food is the most resilient category and most sought-after category for convenient consumption. If you just think about the fact that it is the activity in which we engage with, 3 times a day, 20 to 25 times a week and more than 100 times a month. And then now you include all of the shopping categories and use cases outside of food that DoorDash has been pursuing for the last 5 years, I think you get the setup for some of the markets and the market expansion that you've seen. But look, I mean, a lot of the credit also goes to the fact that we're always trying to make our products better, right? So our product today in '25 is better than our product yesterday in '24, and our product next year will be better than our product this year. So you have this big market out there where we're single-digit percentage of that when you look at the number of occasions that we actually capture today, we still lose the vast majority of those occasions to the pantry or a different form of consumption. And so it tells me that we actually have a very large runway ahead so long as we can keep improving the product. And I think one of the things that tends to get underestimated is how much you can actually keep improving the product. And my take on this is that there's a lot of room. I said this before about our products into other categories, but even our product in the restaurant category, I believe, while it's best-in-class, I think in the eyes of the consumer, it still has a lot of gaps in terms of where we can improve. And that's what we're working on every single day. And that's why perhaps they don't necessarily yield material benefits in the current quarter or even the next quarter. But I do know that if we keep working on this, the compounding does have that effect in the years to come. And so sure, perhaps we've been fortunate to serve lots of customers today. But when I look at the actual occasions that we capture today, whether it's food consumption, retail consumption, and then when I look at this on a global basis, and then when I also include shopping inside stores, not just ordering to the home or to the office, I think we still are very, very, very small and early. And we can't underestimate just how much work we still have left to do to satisfy increasing customer expectations.

Nikhil, it's Ravi. Look, I mean, I think the second part of your question was around like, hey, what's driving the growth. Look, everything is important, right? Size of new cohorts is important, size of existing cohorts is important. Obviously, mathematically, the size of the existing cohorts is larger than new cohorts. What we're seeing is the size of new cohorts is larger today than last year. Size of existing cohorts today is larger than last year, which points to 2 things, right? One is people are coming back more often, which means retention is increasing. All of that is being driven by the underlying improvements that we're making in the product. The second dimension you should think about is order frequency. Again, order frequency hit an all-time high. But again, we are a small fraction of the total number of usable movements just across restaurants. If you add other categories, there's still a lot of opportunity for us to grow. So if you're thinking about it from a model perspective, Nikhil, I would say growth is being driven by both users as well as order frequency. On the user side, both new as well as existing are higher than before.

And Ravi, if I could just follow up with a separate question. The net revenue margin expanded as promised. How are you thinking about that for the back half of the year?

Thanks for saying as promised, Nikhil. I mean, look, we talked about the fact that take rate was going to be higher in Q2. It was in line with our expectation. I mean, look, our business, there's factors that drive the take rate, specifically in Q2, Dasher cost is seasonal for us, right? Again, I'll reiterate, we've talked about this before. Dasher costs are higher for us in Q1. Dasher costs are higher in Q4. So you're thinking about the model, Nikhil, in Q4, Dasher costs are going to be higher. They're more reasonable in Q2 and Q3. So as we came into Q2, we saw benefit from Dasher costs. Two, we are driving improvements in the underlying product, especially quality, which gave us benefit in terms of credits and refunds. The third factor was ads is becoming a larger portion. So if you put all that together, that's what drove the increase in take rate in Q2. Now if you're thinking about the rest of the year, right, Nikhil, I would think that H2, the second half take rate is going to be higher than the first half take rate. But just taking a step back, I mean, look, we are not optimizing the business towards a specific take rate, let alone a specific take rate within a specific quarter. Our goal has been to focus on overall profit dollars. Our goal is to invest flexibly up and down the P&L. That's actually what's driving the strength that you're seeing in the business, right? We have the ability, we have the opportunity. When we see opportunities to invest, we double down. That's what's driving strength both in the top line as well as the bottom line.

And your next question comes from the line of Michael Morton with MoffettNathanson.

I think this is probably for Ravi. I wanted to ask some of the operating expenses. You typically refer to them as fixed operating expenses. They're running maybe a little bit hotter than some of us might have expected. Not that you should manage the business to our expectations, but we're living in a very interesting time where you hear a lot of the tech leaders talk about no headcount growth, utilizing AI within their corporate structure. So I was curious to how you think we should think about your needs for headcount growth going forward? And then maybe just it could be helpful with kind of the fixed cost line item in the model.

Yes. Mike, I'll give you the model part first and then talk about the philosophy. Look, just trying to model, I would say OpEx, roughly, you should think about it as 2% of GOV in that range, which we've talked about for the past couple of quarters. Over long term, like, look, I mean, we think of that as any other part of the P&L, goal is to drive leverage, goal is to drive efficiency. Philosophically, I mean, here's how I think about it, right? Like, Mike, look, we are investing in the business. We're still finding great pockets to continue to drive investment in the business. We're being very disciplined. We've added people, both on the product and the engineering side, in specific areas where the return on investment continues to be great. When you look at the output, I mean, the output is showing up in terms of retention, output is showing up in terms of order frequency. The output is showing up in terms of improvement in unit economics, right? For us, all of that is being driven by the underlying improvements we are making in the product. We still think there is a lot of opportunity for us to continue to invest. But look, I mean, longer term, goal is to drive leverage on this. We are driving a lot of automation inside the business across the board. In general, the way I think about this is the real cost of operating the business and goal is to generate leverage just like any other part of the P&L.

The only thing I'd add to what Ravi said is you got to think of DoorDash as a growing set of businesses, right? So DoorDash 5 years ago was largely one product and operating in one country, restaurant delivery inside the United States in terms of what contributed to our financial results. Today, we have 5 businesses. We have a business outside of the U.S. We have a business outside of U.S. restaurant delivery in all of our category expansion. We have a business with our commerce platform. We have an ads business, and we're working on new businesses. And so I think one of the ways in which we've thought about this is where do we actually deploy our best people and are they actually working on the right problems and do we have the right number of people doing that. And a lot of our headcount growth is really geared towards working on new problems that we see that we can solve for local commerce versus some of the existing ones. That's maybe a bit hard to see in the averages, but that's a bit more what's happening under the hood.

And your next question comes from the line of Andrew Boone with Citizens.

It sounded like Wolt had a pretty good quarter in terms of frequency as well as unit economics. Can you guys just help unpack what are the improvements you guys are making in that business? And then I wanted to go back to frequency. Frequency continues to improve in terms of the U.S. and was at all-time highs. Can you guys unpack that? Is that new verticals? Is that wider cohorts that you guys talked about maybe a year ago of people that order once a quarter and now ordering once a month? Is there anything else you guys can shed in terms of light on why frequency continues to just hit new highs?

Andrew, it's Ravi. I'll take the first one on international. Look, I mean, international business, a really strong quarter. When I think about it from a growth perspective, growth continues to be quite strong, coming from both growth in users as well as order frequency. When I look at the MAUs just for the international business, they have hit an all-time high. A lot of that is being driven by, look, I mean, the underlying improvements we're making in the product, right? We've added more selection. We've added more categories. We have a new verticals business in international that's also growing quite nicely. Wolt+ has been a good addition to our overall portfolio. That's a subscription program internationally. We've launched that about 2 years ago. That continues to do quite well. And when I look at the slope of the Wolt+ curve, actually, it's growing faster than DashPass at the same time. At the same time, the improvements in both quality as well as affordability that's driving people to come back to us and order more. So order frequency has also hit an all-time high. So when I put all of that together, right, our international business, not only is it growing, but we are gaining share across most of the countries that we operate in. We're really proud of the progress that the team has made. And the second point I would make is when I look at the unit economics, I mean, I mentioned that gross profit is positive for overall international, that continues to be the case. And even on a unit economic basis year-over-year, we've seen a good amount of improvement. So net-net, I mean, business is strong, both on the top as well as the bottom line for our overall international portfolio.

Yes. Andrew, it's Tony. I'll take the second question on frequency. The way I've always thought about this is how do we solve more and more customer problems, so that there's more than one way to win, so to speak, where we can actually engage customers that makes the most sense for them in terms of actually increasing their frequency. Like no customer, to put it a different way, thinks about DoorDash's order frequency as a metric that they care about. And the short answer is it is different for different customer groups. There is obviously the introduction of new categories and new use cases. There is the introduction of DashPass and the growing benefits from DashPass. There is also improvements within each one of the use cases, whether it's restaurants. And within restaurants, there's different occasions within meal times, for example. There's different categories, obviously, outside of restaurants. Within each one of those categories, there's various different things beneath that, too, so that you can actually improve each one of these things. So there's no one thing or 2 or 3 things that drives the growth of frequency. And this is, in general, how we set up a lot of our product teams so that, a, it's actually solving for that specific unique customer what it is that actually solves a problem for them so that we can grow a metric we care about and then finding more than one way, ideally several dozen ways, to win so that the growth is not only enduring, but also geometric.

And your next question comes from the line of Youssef Squali with Truist.

So 2 quick questions. One, on SevenRooms, I think that acquisition closed in early or maybe mid-June. I'm assuming, Ravi, that the contribution was de minimis. But maybe just talk about, Tony, how does it unlock new opportunity for you? If you had to dream the dream with kind of building a SaaS model around SevenRooms, can you maybe just share how you kind of see that evolving for you? And then in terms of profitability for new verticals, i.e., non-restaurant, can you just provide an update on how has it progressed as your growth has accelerated? Does it continue to improve on the margin? Or are you holding the profitability relatively flat to maximize top line growth?

Yes, on the first question on SevenRooms, I'll take that one. I mean, look, we started our Commerce Platform business in the fall of '16, 2016, and in earnest in the first quarter of 2017, where we shipped DoorDash Drive. And the dual mission of the company has always been, on the one side, we want to help you grow and bring you incremental sales. That's the job of our marketplace. And on the other side, we want to empower you to do it on your own. So every physical business, small, medium and large, inside cities can win in the digital economy. And we now have built or shipped two of the most successful B2B products in our category with DoorDash Drive and Storefront, each serving hundreds of thousands of businesses. But it still is a very small -- it solves a very small fraction of all the problems that a business needs to solve or even compared to the DoorDash marketplace, it's still a very small fraction of the product portfolio of what we've built for ourselves. What products like SevenRooms bring is really it brings a marketing component and a data analytics component to understand all of the different customers that engage with each one of these merchants in an omnichannel 360 way. And obviously, if you can marry the best-in-class product, which we believe SevenRooms has built, with the know-how as well as the data sets that DoorDash has access to, I think we can build something very remarkable as a third B2B product that will be very, very successful to many, many merchants, both in restaurants and beyond. And so this is less thinking about the business model, although you're right, this is a very different type of business or B2B commerce platform, which contributes to metrics like revenue, but not GOV, but really, the way I think about it is what are all the different problems we can solve. Because if you can solve all the different problems for these physical businesses, I think you're going to grow the GDP of the cities in which these businesses reside. And if you can grow the GDP of the cities, then I think that's a win for everybody. And so that continues to be how we've always thought about this, and it's why we're really, really excited about closing the acquisition of SevenRooms.

Youssef, on SevenRooms, you're right, I mean we are investing in the product. But from a bottom line perspective, you're not going to notice the impact in the second half compared to the rest of the portfolio. To your second point around new verticals, I mean, look, tactically speaking, right, when you purely think about it from a unit economics perspective, yes, unit economics improved in our new vertical business year-over-year. The thing that you have to remember is like, look, we have a structural advantage. We have consumers existing on the platform. We have Dashers existing on the platform. So when you're thinking about it from a modeling perspective, right, like the flow-through from gross profit to contribution is going to be very high whenever we stand up newer categories. I feel very good about where we are on unit economics. What we are focused on is driving scale in the business. Look, ultimately, what we're trying to do is the same playbook, right? We are trying to improve the unit economics, take that, reinvest that back in the business to drive retention as well as order frequency because that ultimately drives scale, scale drives efficiency in the business. And you're seeing that come through in the business, right? When I talked about Q2 to Deepak's question earlier, look, I mean, cohort sizes are increasing, both new as well as existing for new verticals. Order frequency is growing. At the same time, the unit economics are improving. We feel really good about the performance. We feel business is scaling, and we're going to continue to scale that business.

And your next question comes from the line of Jason Helfstein with Oppenheimer.

I wanted to ask a bit more about DashPass approval. I appreciate the disclosure in the quarter. So what has been basically the kind of most effective method for growing DashPass users? Does it differ by country? And then talk about how some of the co-marketing partners have played into the DashPass expansion times.

Jason, it's Tony. I'll take this one. I don't know which earnings call I may have said this before. But look, when it comes to DashPass, the 80 in the 80/20 of DashPass has always been, first and foremost, make the product more useful, make DoorDash, make Wolt more useful. And the more useful we can make it, the more likely that a product like Wolt+ or DashPass will actually get used. We don't want any leakage in the system. We want a product that people see a large multiple in terms of the benefits they receive versus the amount they have to pay for the subscription. That's the 80. That continues to be the 80. There's no one thing that contributes to the 80. And so it's mostly an obsession over focusing on every product detail and increasing the number of benefits. The 20, though, to your point, in my opinion, gets earned after you achieve the 80, which is, if you can achieve the 80 and build the best-in-class products or the most useful ones, and I can't think of any other program that can connect or get used as often, if you can maximize the number of connections between consumers and businesses, then you can achieve the 20, which is to build a successful partner ecosystem. We've been very lucky in that where starting in 2018, 2019, we partnered with Chase. And we haven't done that many of these, but we tend to go deep and partner with a few like-minded partners. More recently, we partnered with T-Mobile, where we get to go super deep in ways that benefit both audiences from both companies. But that really is the 20, and it can only happen if we've actually built the most useful products to start with. And so that continues to be what the focus for both DashPass and Wolt+ are.

And your next question comes from the line of Michael McGovern with Bank of America.

I have 2, maybe one for Tony and one for Ravi. I'm curious on the Coco Robotics partnership for sidewalk robots that I think occurred in April. Any insight into how that launch has gone in L.A. and Chicago or any other new thoughts on robotic delivery methods? And then second, it looks like the AOV or basket size has continued to see this uptick in growth for a couple of quarters in a row now. Anything to call out there in terms of -- or just normal kind of growth in food costs?

Yes. Michael, I'll take the first question on robotics. Our work in robotics started in 2017 when we were really trying to understand how autonomous delivery can actually be shaped and built and ultimately commercialized. And I think the biggest thing that we've learned there is that doing autonomous delivery is actually very different from doing autonomous passenger driving or robotaxis. And that's because, obviously, in a passenger context, the passenger can walk in and walk out of the car even if the drop-off or pickup locations aren't perfect. But in the case of delivery, that's obviously not true. And in order to make sure that you can make autonomous delivery happen and achieve its full potential, you kind of have to solve for the end-to-end system. And that's kind of probably the single biggest learning we've had. And so we've made a few partnerships, as noted in your question. Those have gone great. We definitely see great potential in building that out. But we've also done our own work as well. And so we don't have any announcements to make at the time, but it's something that we've been studying and working on for several years now and something we're very excited about.

Mike, on the second point, I think you're probably looking at for the overall business. I mean, look, largely as a result of mix shift, we are seeing new verticals become a larger portion of the business. Even within new verticals, what we are seeing is an increase in basket sizes as users try to use us for more use cases. A lot of that is being driven by the underlying improvements we've made in the product. Look, we have more selection than ever before. The underlying product continues to get better and more easier to use. That's what's driving the higher baskets in new verticals. But the overall AOV at the total company level is largely a mix shift to new verticals.

And your next question comes from the line of Lee Horowitz with Deutsche Bank.

While we appreciate grocery seems to be a big priority today, can you maybe update us on the state of your retail business? Any indications on the pace of growth here, how the unit economics are trending and how you may be thinking about investing a bit more aggressively in this vertical. And then maybe an update on drone delivery, if you could, some regulatory news in the U.S. making it easier. You guys have been running some tests in Europe as it relates to drone delivery. I guess how do you see that product perhaps changing the unit economics of your business over time? And how do you think about maybe leveraging partners versus building yourself?

Lee, it's Tony. Maybe I'll take both. And feel free to add, Ravi. The first question on retail, I mean, retail is a really exciting business for us. I mean it reminds me a bit of, in where it is today, kind of where grocery was for us maybe in 2021, something like that. Put a different way, we're just at the beginning, and we're just scratching the surface. I think the biggest learning that we've had on retail is that it is just very, very different buying something that is nonperishable and also very different based on each category in terms of the SKU depth or the item coverage that you need to build a product that actually addresses real customer problems as well as everything from redoing the entire shopping experience to the post-checkout experience and all of that. And so all of this to say that it's a sizable business growing for us today, growing super fast, but it is so early in terms of where it is as a product. And so back to one of the comments I made earlier on this call, I think it tends to get underestimated, the number of things you can do to improve a product or the amount of time it takes to actually do it. And then once you do it, the customer says thank you and then they say what's next. And so those expectations always go up. And so we got a long, long, long ways to go in retail. We may be a leader or perhaps even the leader in third-party today in that category. But I would say, as a product, it is a baby in terms of where we are. The second question, I think, was about drones. You're right, we're very excited about some of the recent work that we and others have done with the administration in hopefully greenlighting drone delivery in the U.S. As you kind of noted in your question, other parts of the world actually have kind of come to this development even before the U.S. And we've done a lot of that with partners like Google Wing, for example. We've announced other partnerships as well in other parts of the world. And we're very excited about those partnerships. The way I think about how this plays out, and we'll have more to share later, but it's that you kind of have to think about all the different use cases and you got to solve for the end-to-end system because it's not just the vehicle itself, in this case, whether you're talking about a drone or in the case earlier of the previous question about a land vehicle, those products themselves, as complicated as they are, actually only addresses maybe 20% of the complexity to actually deliver something that I think could be scalable, economically viable and most importantly, actually solve customer problems in a way that human drivers cannot. And so a lot of work needs to be done. We certainly have very strong points of view given that we've been working on this for about 8 years at this point, and we look forward to sharing at some point in the future.

And your next question comes from the line of Brian Nowak with Morgan Stanley.

Tony, you've made so many impressive improvements to the platform, and we can see it in the results. Let me ask you a little bit of a tricky one. Can you talk to us about the areas where you've had more challenges than you expected in sort of improving the product, improving the customer experience, improving the merchant service. What are those areas that are proving to be harder that we should sort of look for the next unlocks to come from long term on the platform?

Yes. Brian, I wish you get to have a peek into my inbox because I get several hundred e-mails a week from all of our audiences, whether it's consumers, Dashers or merchants. And I don't know if they think that our improvements are very impressive. Those e- mails tend to perhaps suggest the opposite, which is a daily reminder, given that I've done customer support now daily for each day for the last 12 years that we seem to be falling short actually. So I'm not so sure that I agree with the premise of the question. It's not tricky at all the question because I actually think we're pretty far behind in each one of the areas from the perspective of the customer, not the perspective of what our business metrics suggest or our competitive position. And so look, I mean, this is true. Maybe we can take an example this morning. I received an e-mail about how we showed up at the wrong parking lot inside an apartment complex. And that caused both a delay in delivery as well as a fear that perhaps the driver wasn't even going to show up. And so whether it's very small things like that, that sounds very small, but when you compound them to the billions of orders that we do, to understanding how do we play now in new geographies or new categories or how might the product change given where the world is going with AI or how business models might be adding complexity to our own business. We talked a bit about SaaS earlier. I would say there are many areas, Brian, in which we're struggling. And every day, I think, is a daily struggle where the job is to try to make an improvement for that day. And one of the things I've learned over the last 12 years of doing this is if you can actually do that, you'd be surprised at both how much progress you can make over a year, but also how systemically you can actually create an environment, a team that actually knows how to repeat this in any new domain, which I think is the most important skill in a world where things are changing ever so fast and frankly, increasingly faster every hour of the day. And that's kind of what I keep stressing to the teams. We kind of make the next improvement. So is there one area that I feel like, "Oh, it's been harder to crack than another." The answer is no because I literally talk to our customers every single day, and it seems like we got opportunities everywhere.