The Founding Story & First Steps For Starting Postmates

You know, if we go back to the very beginning with Postmates, you know, it's a marketplace we of course all know of today. But, you know, what was the founding story and what were the first steps that you actually took to start it as a marketplace at the time? Well, we actually, all of us, the 3 founders, me and one of the founders, Sam, we worked on a slightly different idea called curated.buy, which was sort of like a real-time curation platform is what we called it at the time. So it, The thought was that certain people, maybe it's like Twitter lists, but instead of putting people in there, you would put content in there and people could subscribe to that. I think we called that bundles. Anyway, it was as stupid as it sounds and, but there were a few people using it who were like media influencers at the time and that got us some press. But we met, we met this other guy, Sean, who obviously became my co-founder and is one of my best friends. And he worked at, at actually a really interesting startup called Hug Energy. And they were trying to build the world's most intelligent power strip. It would meter the consumption of devices attached to it and it would help you turn things on and off in your house. Now, this is probably before, like, life science became cool. I think with the Nest thermostat, people got the hang of it. This is, this was one of these products that was, was too early and, and, and, and people didn't really care about the, the consumption of their devices much. So we, we started working on this idea. that I had a while back, which already had the name for Postmates. And it evolved quite a bit from what it originally was. I think the original idea was, can we sort of create like a local postal service that allows you to move things within the city? And turns out, you know, There, you needed a sort of catalyst for people to have the, the original use cases were like, what if you lost your keys or someone finds them and you wanna get them to you? But, but I think pretty soon and, and some other iterations were around, we, we wrote this, this really beautiful software actually that, that would do that. So it was, it was, it was one of the first iPad, iPhone apps. And it looked really great. It looked like a notepad, you know, just like as if you were to— skeuomorphic design was still huge then. And, and so you could type in the pickup address and the drop-off address, and you could enter a description and you could upload a photo. So, and now we're, we're, we were here sitting and we were like, okay, clearly now everybody is is using this to get things delivered through San Francisco. And when that didn't happen initially, we signed a lot of retail stores. So we went to retail stores and we said like, look, do you ever have people coming in here and they asked if you deliver? And we, they, many of them were like, yes, this just happened a minute ago and I couldn't sell this chair because this person doesn't have a car. And And we were like, well, yeah, use Postmates the next time someone wants to buy something and have it delivered. All you need to do is put the information in and we'll deliver it for you. And we had like 20 stores signed up. There was a vintage store in San Francisco called Stuff on Valencia Street. There were Els Attire, which is a tailor in North Beach. Really just great retail stores. But the volume and there was a bike company and the volume was very low. But what happened somehow is that, you know, in this app that I just described that really had no sort of payment functionality whatsoever because it didn't really exist. I don't think Stripe had launched yet or They actually hadn't launched. So what would happen is that in the description, well, at the pickup, people would put in a store, a restaurant or Whole Foods, and in the description they would put the things that they wanted us to get to them. And we had to cancel all these requests because we were like, you guys, this is not how this is supposed to work. And also, So we were actually annoyed by it because we didn't want it to be like a company that does that sort of thing. We, we thought about us a bit more as a, we wanna build an infrastructure that is like used for commerce, not be the errands guy, right? And so, but, but we talked to some of the friends and I think Jack Abrahams and a few other people, they suggested, look, maybe give this a try. And I, and, and we were very reluctant because we said, how can we give this a try? Because people order an order at a restaurant. What do you want us to do? And we ultimately came to the conclusion, what if we just call and place the order and find a way to pay or have someone pay for it? So now again, there is no, we're not, there's no integration done there. By the way, there is no, there This is in 2010. There is no point of sales yet that have APIs that you could, even if you had an agreement, could hook into. A lot of places are cash in San Francisco especially. So we did 2 things. We, we bought Visa gift cards and these Visa gift cards, they get them at Safeway. And we would hand them to the fleet of couriers. We had around 25 couriers in San Francisco. 10 years later, obviously you have, you know, Ben Horowitz writes about do things that don't scale, right? So this is before he even wrote a book. So the, so we gave these gift cards to the fleet. Now they're as good as cash. So you have this group of people that is not employed. And we wrote a little software that would frequently call the, there's a, there's a number on the back of the card and you can, you can type in a code and it tells you the balance. So we wrote a little software that would auto dial this number and check the balances of these gift cards. A, so that we don't get ripped off and B, so that we could keep track of the spending. We also asked the fleet to write it down on like a little A5 envelope that they had that we provided. So we, we try to keep a tally of that. Now, now they had a way to pay. They could also get cash. So we would take cash out and we were Silicon Valley Bank and they, why do you need money? Why do you need cash? And this is so funny, like, you know, you, you get some funding, but nobody wants you to touch that money. So we were like, well, we just need it. So we ran a, we ran some pocket cash as well, or like a telly for that because a lot of the places that were, we got requests from, a lot of the burrito places, Mexican food places back in the days that were El Farolito was cash only. And I don't know how many people of you live in San Francisco, but so we're We're now set to purchase things. But we obviously have no way to charge the customer because again, the rails of just dropping in an API key and having an interface didn't exist via Stripe. So we bought a few dozen Square readers and we set up a Square account, founders@postmates.com. And we would instruct our fleet that they take a photo of the receipt when they purchase the item, and their fleet app would show them the amount they have to punch in at the Square reader. And the loop was closed, right? So a, a, now the only thing we had to figure out is how do you get people to order this frequently to make it a real test? So we, we couldn't just route every order to a Postmate, which is what we call the fleet, because only certain people had a, had a Visa gift card or cash. So we tagged the couriers in our system with a hashtag, get it now. And we Wrote an email and I still have this email somewhere. And it says, introduce— we had around, there were a few thousand people on that, that had downloaded the app. Again, we got lucky because some of the tech press really liked Curated By. And so when we then pivoted to Postmates, they wrote about this and a lot of people then tried out Postmates. And had the app installed that was pretty useless but looked good at the time, I guess. So we wrote an email and that email said, this weekend, if you use the hashtag #GetItNow in the description in that little field, we go to a store and pick up the things that you want us to pick up and we buy them for you. No, no questions asked. And we sent that email on a Friday. And by Sunday evening, I think we had sold almost $10,000 worth of goods from stores. It was that dramatic change from one thing to the other. Now, and with that, the wheels came off. Because people wanted it, but every second order was wrong. The couriers didn't understand what to order, didn't read the description correctly, didn't know what to do. Long wait lines at restaurants and nothing was optimized. But the one thing that was clear was that there was consumer demand and people that used it that weekend, they tweeted about it and They got In-N-Out or they got El Farolito and they couldn't believe it because what is extremely important to understand is that there was no other service like this in the entire United States easily. But in San Francisco, even what was in place was so inferior to what one would be used to in New York when you had a guy the liquor store would actually come upstairs. If you live in the Village, you'd get someone to do these things, even though with a lot of effort, you, New Yorkers had a different lifestyle. But in San Francisco, you had Inferno Pizza and maybe another place and that was it. So people just couldn't believe it. And the only other thing that I say is we had to solve a third thing then in order to scale. And that's what we call the board. Now you have to understand, up to this point, all of these orders were dispatched manually because what do you need to write software for when you have 25 people? So the software that we use is they had devices, we tracked the devices on a screen and we would see where they're going to depending on, on what they had accepted. But there were no routes and there was obviously no intelligence on routing anybody. So we had to manually run this board. And you actually can't write routing software from scratch for just 25 or 30 moving things. There's no efficiencies you get from that. By the time you're done writing it, you don't even know what you did it in the first place because there's just not enough data points. So you have to run this system manually for a while. And what that meant is keeping an eye on the courier or the Postmate. And we also started calling in these orders. at the restaurants and said someone is gonna come and pick it up. Mike's gonna pick it up. Whoever is gonna pick it up. Now that, that from the, this, from that weekend on for about 2 years became a full-time job, a full-time job every day from 9 in the morning till midnight for one of the founders. What we did is one of us would run the board. Especially after hours. After a while, we had a customer service person run the board during when we had to work. But when one of us wanted to go home, either Sean or me, I, one of us would go home, have lunch, have dinner, would let the other person know that they took control of the board. And then that person was on call from 7 o'clock until 11:30 or whenever the last orders in restaurants were accepted. To place this order in, let the courier know that it's been ordered. Hundreds every day and soon thousands every day. So it was an extremely stressful job to do that. But again, there was no infrastructure to relay these orders, and we had obviously no agreements with any of the restaurants. And that's maybe the last thing that I say, and this covers actually more than just one thing. This covers how we got it done. The beauty of Postmates and its total complexity came from the same thing. And that is that we allowed you to order from any place possible. We didn't restrict you to order from just a few restaurants we were able to sign up because would've been so few that the whole magic of the product would be extremely diluted. Now, would have the— would the experience have been better if we had waited until we have agreements and proper infrastructure in place to order from a handful of restaurants for these restaurants? Absolutely. But we made the decision that the overall magic and that feeling that you can get whatever you want without virtually any restriction was the far more important thing to do. Even if that meant 2 or 3 years of pain, actually a lot longer. We can go into the details why, but a pain, but ultimately creating a product that, that where you, where you, where we could use the tagline, get anything from anywhere. Which, which was, which was, which was amazing. I stop here.