How Thumbtack Has Evolved As A Marketplace

you walk us through how Thumbtack has evolved over the years and, you know, kind of what led to each of these evolutions and maybe even some of the learnings?

I'd say there's a continuum in some sense, but there's like the dark ages, which is like before we had our V1 of like the first, you know, end-to-end request for quote solution. That's what sort of Thumbtack originally was. That took kind of maybe 3 years to fully put together in the sense of like a pro side that attracted pros, a customer side that attracted customers, sort of matching and connection sort of solution in between and a way to monetize that. You know, we launched, we started work August of '08, 3 weeks before Lehman Brothers goes under. And that V1 solution was in place At the like very start of very like middle of 12, right? So even 4 years to get to that true 1.0 solution. And that's really when, you know, Thumbtack became began to scale. And you know, as this request for quote offering, what we had offered and built for folks was, you know, a way for customers to avoid. Than a sort of drudgery of calling down a list of names and numbers, right? Very easy to go on Yelp or on Google and get a bunch of names and numbers, but then you were stuck yourself just calling and that was annoying. And so our system let the customer sort of specify once what they were looking for, and then we use that to sort of send it off to the pros who are relevant and interested in that type of work, who would then choose to pay Thumbtack for the privilege of responding to that customer. which meant that they were kind of self-qualified and really kind of high-value matches. But that— and that helped us scale because it was better than directories, it was better than classifieds, it was better than kind of the alternative that was out there. But the reality was that it was still sort of highly imperfect. And, you know, the tension was we could get you these high-fidelity matches, But it was because we had a search engine that ran through the skulls of human beings. We're very good at matching and qualifying. And when you got something, it was really useful. But sometimes they were tired or under a house or on top of a house or busy or whatever it may be, and they missed it or didn't respond or didn't have it in them to sort of send that message. You, the customer, were deprived Kind of a plausible match that maybe could have been a great pro and not ready to sort of respond to the request. And so it was kind of limiting because, you know, here we were at the end of '17 and into '18, and it was clear that this model could keep getting better, but was sort of fundamentally kind of limited in the sense that, you know, if it wasn't instant, if you weren't getting enough responses and You know, we struggled to get liquidity above kind of 2 to 3 quotes per request, which is like a lot relative to 0 or 1, which is what most people get, but far from e-commerce-like experience of being able to kind of browse, shop, you know, change, go back, you know, filter, et cetera. And so then we made the call of like trying to innovate our way out of it. And This— the way to describe it is we wanted to automate what the pro was doing in their head. So retain that sort of matching sort of algorithm and criteria that they were expressing so that the matches could be as intentful and as qualified, but do so programmatically so it could happen instantly and you could get search results. And we called them sort of instant quotes or instant matches rather than this asynchronous call and response system, which Yeah, just wasn't meeting customers' expectations. So it really was kind of a wholesale change of everything, the customer experience, the pro experience, how the matching models, the business model in some sense. But the single hardest part was it changed how pros paid us and what they had sort of control over. So in the old model, they held their credit card and they swiped it every time they wanted to respond to a customer. Very clear, very, doesn't take much trust in the sense that you just read, pay, and respond. In this new model, we were saying, hey, look, you tell us the types of customers you want, we're only gonna bring you those. And for those, we're gonna charge you kind of a sort of preset and then ultimately kind of floating and dynamic amount up into your weekly maximum. So we swipe your credit card whenever a customer comes through. And that sort of change of going from like, I control to I have to trust, and this is more like Google or Facebook, and I have kind of a weekly max and a cap, and then I specify what I want, was super hard and took, you know, kind of 3 years from the— All of '18, and it wasn't till kind of '20 That we kind of finished that transition. And that ultimately kind of powered the next leg of growth for the business. And, you know, interest rate changes notwithstanding, is sort of still powering that growth to this day. Because ultimately the truism and something that I think all of you can trust applies to your own marketplaces is, you know, Consistency of fulfillment, right? How often the customer gets to done is going to be a core driver of their loyalty and retention forever. You know, you look at like, how did DoorDash win? You know, I think it's just a relentless focus on fulfillment experience and continuing to just grind up that, you know, in their case, I don't know what they referred to it. We call it our higher rate. Right? How often the customer's hiring, they're probably thinking of the buy rate or things like that. But point being, marketplaces and, you know, stores and e-commerce, however you want to think about this kind of category of solution, is really about providing sort of ease of use when it comes to finding and buying what you want. And the more efficiently you can do that, typically the better loyalty you can drive on the demand side and ultimately the better sort of retention you can have of your suppliers who are interested in sort of being bought, broadly defined and sort of selling more. And that's hard and it takes kind of a relentless focus and it is kind of myopic because there's all this other stuff you could go do around it. But that is the core for most marketplaces. And typically I would say most fail because they fail to create enough liquidity to really satisfy that core desire of wanting to find and buy something.