Learnings From Scaling Gigster As A Marketplace For Software Development Services

You know, going back to Gigster though, could you maybe share a little bit more about, you know, what you were working on and kind of building with it as a marketplace at the time?

Yeah, certainly. So we were, you know, as a startup founder, you're often a customer of like a software development agency. You know, many of us on this call have probably hired developers online or hired agencies online, and I never really liked the experience. So my co-founder and I thought, how can we make that better? And there's obviously a great product there for SMBs, competing with the Upwork or oDesk or Elance of the world at the time. But there's also a great product there for enterprises. They have to work with Accenture, Deloitte, these massive firms that famously, I think Accenture charged the TSA like $1.5 million for that app with an arrow that would point people left or right. You know, so we did a little viral thing after that where we're like, hey, we can build— we built that app in 30 minutes. But anyway, so it seemed like an area ripe for disruption, and our approach was this, was to basically sell the customer a kind of a project and not time. So talk upfront, get aligned on what they want, and sell them a project-based fee. So we'll make your entire app for $15,000, for example. So that was one major piece of it, which is more customer-friendly because you know upfront what your costs are going to be. 2 was we had a highly curated marketplace and everybody says that, but we were literally getting our friends that worked at SpaceX and Google and Airbnb and Amazon. We were getting our friends on the service and these are people that you really had no other way to hire anywhere else. So we really truly had some of the top software engineers in the world. We also, rather than giving you an individual person, we gave you a team. So it wasn't a marketplace of individuals. We were the marketplace orchestrator, and we would give you one product manager, we'd give you maybe 2 backend engineers, one frontend engineer, and even a designer. So we gave you all the people you needed to, to get your complete project done. So kind of, you know, several marketplaces combined into one to deliver that final product. And the last part of it I would say is we, at the time, AI was having a renaissance as this kind of popular buzzword and term, and machine learning was starting to be a thing. And we had played around with a little bit of that. So we wanted to give our marketplace members kind of Iron Man suits, in a sense, is what we called it, where everybody in the marketplace had these augmented superhuman abilities where they could be better and make more money and do more for our customers. So we really believed in kind of building, using AI to augment these marketplace actors, which at the time I think was a little bit too early and we didn't fund it enough. But now you're seeing more and more companies like that. So yeah, a couple innovations were tied into it, but we were trying to, at a high level, just build a better software services marketplace.

Yeah, yeah, that's really interesting. And thanks for sharing more with us. So definitely I remember seeing the Gigster, you know, the name back in the day. So, you know, kind of, I would say like looking back now, you know, what were some of the biggest challenges with it as a marketplace specifically? And then maybe, you know, what are some of those kind of learnings that you've then taken with you to Maid Renovation?

Yeah, good question. So I touched on a couple of the things we were doing and I'll tell you which ones worked and didn't. So the whole idea of doing kind of project-based versus time-based was very challenging to make work. You know, customers would kind of hold us hostage and, you know, if we were selling an app for $15,000, we would do our math on the backend to ensure we were gonna make a profit. But then as we delivered that project, it was very hard in practice to make the margins work out because customers kind of wouldn't let us go. They keep asking for changes. We'd have to then go after the freelancers and make them do the changes. So project-based versus time-based was probably the wrong idea, or at least we didn't figure out how to do it. I think we probably could have figured it out if we didn't focus on growth over proving the model and proving kind of the unit economics. But we immediately started growing before we really had the unit economics proven. So that was yet another mistake. The kind of Iron Man suit AI augmentation thing didn't work. Part of it is our scope was too broad. So this is another thing with a marketplace is it's helpful to find a niche, especially if you're trying to build a management layer or any type of augmentation element to it. You want to be more narrow rather than broad. The more narrow you are, the more likely it is you can build meaningful technology and tools that actually move the needle for a transaction. So we didn't manage to build any of the AI that we talked about and told investors we could build. And yeah, I think just in general, I would— guidance I would give is sell time and not outcomes if you're doing a services marketplace, and try to be more narrow or specialized.