00:25:19 · Aditya Agarwal
[00:25:19] S1: Definitely have... It's been a pretty good aid to help me raise my three children. Absolutely. Maybe to move on to how, you know, instead of focusing just on AI, but kind of his company building beliefs, what's maybe one cultural aspect of DoorDash that both was present from the super early days and kinda continues to be a north star today. Obviously, we have talked about customer obsession. Are there one or two other ones that are really, I would say, foundational to the company DNA?
00:25:48 · Tony Xu
[00:25:48] S0: You know, DoorDash has... Sits at this, like, intersection, if you will, of of, you know, almost like physical operations that have to happen, and you have to build technology to scale it. Otherwise, there's no way you can do billions of these, you know, you know, transactions, and both sentences are correct. And, you know, one of the things that we've learned early on is that it is very hard to sit at your desk or have two really smart people argue with the most compelling data, to opposing points of view on how to solve one problem. The best way to actually collect information or action is by shipping something. There's this value at DoorDash called bias for action where, you know, the way we settle debates is we ship something. It doesn't mean it's the correct thing. Almost always, it's a hacked together prototype, maybe not even a prototype. Maybe it's, you know, just calling up a customer and trying something with them that ultimately gets us more information. It doesn't mean that it's perfect. It doesn't mean that it's a perfectly statistically, you know, strong answer. But that's been something that, you know, I've always been very proud of at DoorDash, where even to this day, you know, we we try our best to answer questions through shipping things versus who writes the best document, who articulates their point of view the best in a meeting, whose title, you know, is the most senior, things like that. You know, another one that comes to mind from the very earliest days, you talked... You referenced that Stanford football example where we're late on every order and we ultimately refund all the customers, is there's this notion that... There's a saying internally at this company where we say one team, one fight, where we don't care that much about what your job description or your function that you you you were recruited into, you know, says about what your job should be. Your job is to help the customer win. That's your job. If that means you should be doing deliveries or doing customer support, regardless if you're an engineer or whatever you might be, you know, recruited into the company doing, that's what you ultimately do. And I think it goes back to pointing out that many different teams are required to solve one very simple problem.
00:28:03 · Tony Xu, Aditya Agarwal
Even though the customer might be just bringing you, I don't know, your your groceries or something, Many teams, not just the technical teams, were responsible for that one order. And it promotes not only this understanding that, you know, there's an end to end workflow that that that that is really important to understand every single component part, it actually promotes a culture where there's very high accountability but low blame, especially when things go wrong. And, you know, that's something that's very important to keep, whether you're four people in my apartment when we started the company or you're now 10,000 people plus. [00:28:41] S1: Is there anything that you've had to consciously almost, like, affect the change in? There was something that perhaps worked for the company in the early days, but as the company grew, actually turns out that in order to have a long term company, this one thing that we all believed in, you see that you kind of had to change your mind on, had to update your priors on.
00:29:00 · Tony Xu
[00:29:00] S0: No. Every year, we ask ourselves the question, you know, what would we do differently if we started the company from scratch today? So DoorDash is heading in... Into its thirteenth year, so we've thought about this about a dozen times at this point. And the short answer to your question, Aditya, is almost every part of the company, there's something we would do differently. Right? And just by asking the question, you're probably gonna come up with some set of answers. One of the things I love doing is really going to, you know, the edges of distributions to study, like, what's what's interesting. You know, for example, I care a lot about what new people say, new users say about our experience. I sometimes will delete my own account and try signing up for the new user flow and see if it's as easy to use as it once was, you know, many years ago. I care a lot about what new teammates say about working at the company. I care a lot also about the opposite end of the spectrum, the power users. You know, the best notes I get are actually from Dashers. These are the drivers on our platform that would write me thousand word plus emails about things that they've observed, and they have tens of thousands of delivery data points that kind of are really compelling, and I learn, you know, a tremendous amount from, and those always update my priors. So I don't think there's one, like, meta point or or or belief that that that that I've changed because in part, I think part of what you're trying to do as the, you know, as the builder is you're constantly trying to update your priors. You're just trying to do a better job, you know, today than you did yesterday and tomorrow versus today. That you're you're... You should be trying to do that in every part of your company. Why wouldn't you? And then so if somebody in the company had a better idea for how to do something, I am always all ears. Yeah. That's a
00:30:47 · Aditya Agarwal, Tony Xu
[00:30:47] S1: pretty interesting cohort. If you think about the set of people who have done the top 1% of all kind of DoorDash deliveries and the amount that they know about kind of the implicit knowledge they might have about our logistics networks, our traffic networks, that's a pretty interesting data set. Yeah. [00:31:04] S0: Absolutely. Absolutely. Tony, you talk to [00:31:08] S1: a lot of CEOs. Right? Like, you're on Meta's board now as well. You've been there for three years or something. Have you found that you believe certain things about company building that are just kind of, like, weird that no no other CEO tends to believe? Then no other CEO Do you have any, like, heretical weird ideas about company building that seem to be a little bit outside of the standard? I guess, you know, two standard deviations away even in weird Silicon Valley.
00:31:33 · Tony Xu
[00:31:33] S0: Well, one of the things that I don't know if this is popular in Silicon Valley or not is that I think companies... And, you know, Meta's a great example. There are many others I can give you, that those companies don't get enough credit for building the core business, which allows them to invent the future and invest aggressively into building the future. I think a lot of attention, right, think about what you read about or what is written about, is always about some shiny object and, you know, kind of the twists and the turns and the ups and the downs. And trust me, they they all... There always will be those situations. But what I think is really interesting that maybe doesn't get as much attention or certainly as much praise is the constant grind of trying to improve a core business. I mean, you think about the core businesses of Meta or any of the large tech companies, they've actually changed a lot since inception. I mean, you worked on some of them, so you you definitely know. I think the constant work to make those businesses better, they don't just become better, by the way. The natural state of technology is that it decays. And so in order for you to attract that new user, you better work on the next thing, even if it's your old thing. And and that's not sexy, and that's and and it's never interesting to cover those things, but I think one of the things I've learned about how you build really large successful companies like the one Mhmm. Like the one that you've worked for is you have to build the core, and you have to build the new. It can't just always be about the shiny objects, and you have to have, you know, almost value systems and systems of almost appreciation as well as, you know... And I don't mean, like, in words, but in expressed action internally inside the company to actually make it equally valuable to work on the core and to work on the new. I think that's very hard to do. I think you have to always do that because you're constantly updating your priors. We live in the world of technology where, you know, the constant is always gonna be change, more competition, and faster iteration. But you almost have to, you know, remember that you have to keep working on the thing that got you there too, and you have to keep doing that over and again.