A Product Market Fit Show | Startup Podcast for Founders
00:30:11 · Pablo Srugo, Chris Slowe
[00:30:11] S1: People listening right now, they just hate you. [00:30:15] S0: I know, but it's again, it's one of those things where we didn't realize the time that was an unusual thing. The fact that we had to work really hard on the revenue side, it wasn't until the modern era where we really started doubling down in revenue and the last basically 2015, 2016 on was a really had to go hard at it and figure out how to solve that particular problem. For the early days, the problem was never the users. If anything, there were days where we're like, "Oh, cut, can they just stop coming for one day so we can clean up this database?" I mean, great. I just want to take the whole site down time; I need to run a full vacuum, like just let the vacuum finish. [00:30:57] S1: Tell me the – so then you shift. In 2010, you and Steve decide to go off and do Hipmunk. What's the story there?
00:30:57 · Pablo Srugo, Chris Slowe
[00:30:57] S1: Tell me the – so then you shift. In 2010, you and Steve decide to go off and do Hipmunk. What's the story there? [00:31:11] S0: Similar story, which was that I never come immediately but I come eventually. I come quickly. It took a little away time. Steve left in end of 2009 and then went off and came with an idea with a friend of ours, Adam Goldstein, to build a travel meta search site called Hipmunk and I ended up joining about a year later end of 2010. I would say that to go back to my point about product market fit, part of the design choice around Hipmunk was let's make a very nicely designed site and product that has immediate ways to get to revenue and that we just have to deal with – let's find a path where we don't have to deal with complaining users and we can get to revenue and really nice product. We succeeded in all directions and then realized that of course that the hard part is getting people to actually come to your site is very difficult. Steve and I talked about this in Retrospect, it's like we just had this notion that we were going to turn it on and people would just come. That has been the story with Reddit. If anything, we built it with notions of scale that are just hilarious in Retrospect of like, oh yeah, we like a couple of architects who are used to dealing with huge scale in Reddit. We'll just make this thing fast and really scalable and easy to scale up, and scaling was never the problem. It was getting users to come. If you could get somebody to come back twice in a year, that was a rampant success.
00:32:34 · Pablo Srugo, Chris Slowe
[00:32:34] S1: It's so funny how life does that to you. You learn all these things, like this time I'm going to get it right and then you forget that. What do you say? [00:32:40] S0: Yeah, and the side effect even was we ended up building a lot of systems that looked very similar to the shape of the technology at Reddit because same architects were building it, which meant that when I eventually came back or we both came back, Steve came back in in '15 and I came back to Reddit in '16, it was almost like coming back to an alternate history where they had chosen a slightly different path because we had actually diverged in our direction. Came back and found that it's like, "Oh, wow, they're still using the first version of this thing that we wrote. We've already written three other versions but that was a different company. Crap." How does this work again? [00:33:13] S1: What was Hipmunk? Was it like a hotels.com, like a booking.com thing or something else?
00:33:18 · Chris Slowe
[00:33:18] S0: I think that the closer analogy would be it turns out that there's two different – oh God! These are things that I didn't know I'd ever have to learn about. It turns out that there are two things that exist for searching across multiple flight or hotel providers. There's something called an OTA, which is an online travel agent, and then there's something called Meta Search. OTA actually legitimately has a travel agent's license. That industry is weirdly regulated. There's actually a whole classification system. Meta Search is you're built as a search provider on top of other providers. We built Hipmunk initially as Meta Search. We made our own deals with airlines, made our deals with Meta Search providers, and the idea was to become effectively an aggregator. I think that the clearest analogy I would give would be, if you go to Google now and you search for flights, they'll give you that overview. It's basically Google Flights, which came out again, came a couple years into Hipmunk existing where they decided to start making a foray into that. After they bought one of the large data providers for for flight data called ITA, they started building out that product. We initially started with flights, and then we moved into hotels because that's the path that everyone takes. The reason why everyone takes that path is because it turns out that hotels are much easier to monetize than flights. Flights you're making maybe a few dollars a ticket, whereas hotels usually get a percentage.
00:34:32 · Pablo Srugo
[00:34:32] S1: How did you end up solving it, the core user problem, which by the way is frankly, demand is most startups problems. That is just how to get users until you get product market fit and then you have other problems, but that's usually the zero to one issue.
00:34:44 · Chris Slowe
[00:34:44] S0: Yeah, I think that Hipmunk's real differentiator was its flight product where we really started. We wanted to provide a way to solve our own problem first which was, it's just a pain to figure out which flight to take. I have a couple of very simple requirements most of the time for looking for a flight, even to this day. I travel a lot now and I miss Hipmunk's UI. We showed flights as a bunch of Gantt charts effectively, like here's a time bar that shows you when the flight is and you can go through and you can basically say like, "I want to leave after 9:00 in the morning and I want to arrive before this time. Show me everything that fits that and they're like, oh look, this one has a really stupid eight hour layover. Kick that out." We created a simple sort called agony, which was you could start by lowest agony and we give you effectively direct flights and flights that had reasonable layovers. Reasonable layovers is what you would expect. 30 minute layover, that's not a real layover because the flight's going to probably be you have to run between terminals, as opposed to, yeah, about two hours about the sweet spot. We built out a quick little bit of mass that showed that with an ICUI. The product really was the UI. It was quickly being able to get to a flight that's reasonable and be able to pick your return path, be able to break down by price, just all the things you'd expect in a product with a very simple visual aesthetic.