Aziz's team solved the cold-start problem by manually fulfilling orders themselves. They found urgent demand during COVID, matching customers to drivers by hand, then used those early transactions to understand how to build a real platform. Introductions to former Uber data scientist Kevin Novak provided crucial mentorship.
Chapters
Aziz, could you talk a little bit about like, what is the first, you know, kind of, you know, I don't know what the right unit is, right? But like, what's— how do you— how did you get your first batch of drivers? What is the first batch of apps, you know, that stimulate the demand side? Like, tell us kind of the early kind of year 1 cold start problem kind of, you know, thing.
Yeah, absolutely. So when Mahmoud and I started working on this problem, I think one of the things— one of the first things we used to do is be very vocal about what we're working on. And anyone who cared to listen to what we're trying to do, we talked about. And that actually created a really nice kind of effect of us closely getting closer to more experts in the space that potentially even had a lot of experience in verticals that we were interested in. So we had a lot of really amazing kind of conversations earlier on. And part of in that early phase was talking to a lot of customers. So we did outreach. We talked to a bunch of customers who were having these problems. It's kind of a hair on fire problem. People would get on the phone immediately from the first kind of email. That was a really large positive signal for us. And in that first batch of customers, I distinctly remember one of them was actually COVID kits delivery. So around that time, it was COVID. And there's this thing where they're saying people are stuck at home because they think they have COVID, so they can't come pick it up. And it'll take 3 days if I ship it, but how do I get it to their house and then have someone pick it up and back to me? And this is one of the earliest use cases where we onboarded this demand for some part of it. Maybe some of us had to get in their cars and fulfill their own. But maybe. But You have demand and you go immediately to supply and say, hey, I have work for you. And that pitch went really well where you say, can you go and fulfill this work for me? Now, remember the technology, the platform, the enablement, this is all very, very weak. None of it, maybe a lot of it nonexistent too. But this is really that starting point where get initial demand, go find supply to fulfill it, get this transaction multiple times. And this is when we start to figure out how to fit this into a platform play. And part of creating these really nice early conversations, I think that's— we had really awesome conversations with Kevin Novak who was very early at Uber. He led their data science team. Kevin in turn introduced us to Andrew as well as an expert in the space. And this definitely was a big part of our kind of early start conversations.
