Matthew: Who are you co-founders? What qualities or skills were you seeking in a co-founder, and how did you know they would be a good fit? Tom: The two co-founders are Chris Wanstrath, and PJ Hyett. I met them both pretty much through the Ruby community. Hung out with them at the Ruby meetups, and at conferences and things of that sort. So all programming centric types of get-togethers. PJ and Chris had this blog called Air The Blog, which was super popular in Ruby community. They were kind of the dynamic duo of Ruby innovation for Rails and their blog just laid out new ways to do things. Better ways to do things, and so they were well known in the community. It was nice to get to know them, and I knew their code already because I could see it. That leads back to the question of open source. Which is, once you can see someone’s code out in the open, that’s certainly a great way to evaluate them technically. As an aside, that’s still how we do hiring at GitHub, is looking at peoples open source and then interviewing them separately not about technical stuff. Because we don’t need to, we’ve already see their code. As far as Chris and PJ go, I think I got along with them at the meet ups and the drinking afterwards. That has always been important for us, the drinking. Well not so much the drinking, but what goes along with the drinking, which is hanging out people, and getting excited about, and talking about technology for hours on end. That’s where you can really tell if you’re going to get along with someone, is when you’re having a couple of beers and you’re just talking about whatever comes to mind. So it was through there that I got to know them and became comfortable with them and could evaluate them in a sense of what they were good at, what they liked to do, and were they good fits personality wise. That made talking to Chris about the idea, and we were both sharing similar ideas all the time, coming together and saying, hey, let’s do this it was very natural. Like hey, I have an idea, I know that you have similar ideas, let’s combine our talent. You do the rail stuff, and I’ll do the front-end and the back-end stuff, because I’d done graphic design quite a bit previously. As well as systems back-end stuff. And so it was just a natural fit from the very beginning, and it’s worked out extremely well. Matthew: From idea to product launch how long did it take, and when did you actually launch GitHub? Tom: I would say it was about three months from the time that we first talked about it, to when we started letting other people into a private beta. So it was about three months before we has enough to say hey, come check this out. And really all it was at the time was user accounts, and get hosting, and you could browse the source, and browse the commits, and that was about it. You could follow users as well, and you had a dashboard where you could see what was going on with the users that you were following, and projects that you were following. So it was tiny, it was nothing, it was almost nothing. But it was enough to see that this had value, because now you could see peoples repositories, clone them. You had forking as well, so you could fork them into your own account, clone them down locally, work on them, push them back up, and then at the time we had [pole] requests, which were more like messages, like hey, I did some work, go check it out. So that’s all we had, that was enough to get people understanding that Git could be used for better types of collaboration. It didn’t really take much, it just took access to those repositories, and some ability to discover and follow the projects you were working on, so that was three months. In another three months, we did our public launch, and that was in April of 2008. The public launch included a bunch more features and tons of bug fixes and the network graph. I don’t even remember anymore, it was a fair bit more polished, but still quite minimal. That’s also when we started charging. So three months till private beta. Six months-ish to public launch. Matthew: Are there any unique metrics or social proof about GitHub that you’d like to share with the audience? Tom: Well, I could talk about a few things. We have 740,000 users signed up right now. We have 2.1 almost million repositories that we host. And we’re on board in users in a rate of about 2,000 to 3,000 per day. And projects at a rate of around 5,000 to 6,000. Matthew: Wow, that’s amazing. Tom: That includes the Gist, which is a lightweight repository for doing paste type things. The way that we get paid is through private repositories. People pay for a private repository, or a set of private repositories and team features and that’s where we make all of our money. That is growing quite well. Also it affords us the things that we’ve become accustomed to, like this nice office. And it also means that we’ve never taken venture capital. We never raised a dime of outside funding. Matthew: That’s amazing, thank you for sharing. We know founders face unique challenges when they decide to start a company. What was the hardest part about starting GitHub, and how did you overcome this obstacle? Tom: I think it was just figuring out what to do. I mean honestly it was a side project, so it wasn’t that stressful. We were working on it while we had full time jobs, so we weren’t running out of money. We were sort of casually working on it on nights and weekends. Well I shouldn’t say casually, we were working our asses off on it nights and weekends. We’d come together every week in person, mostly we worked online and used IFC to communicate. Figuring out some of the really important things like pricing, how were we going to price it? Were we going to charge for it at all? If we did charge, were we going to charge for open source stuff, or only commercial stuff. While those seemed obvious in retrospect as things always do. At the time, we had very long conversations about it. As far as what the pricing should be the kind of things where you don’t have information and so you’re just guessing mostly, making your best guess, hoping that it works out. The nice thing about the web is that you can change those guesses rather easily. But honestly, in the beginning it felt very natural to do. We were building it ourselves, we were building a product that we wanted to use, and we were building it because we wanted it. That made a huge difference in the complexity of doing it. Because we didn’t have to go and ask potential customers what they wanted we just knew. We were just building what we wanted, and hoping that developers would like us. Things get more complicated once you start hiring a lot of people, and you have enterprise customers, and you have people who fear change. And they freak out any time you change any kind of interface element. Those are the things in the long run are more complicated and difficult to deal with. Matthew: Since you’ve been in operation what have you learned about your business and your users that you didn’t realize before? Tom: I think one of the really big things that we learned was when we put out the organizations feature, which was features for teams. We put them out at price point that was higher then people were used to because we said, we’ve created all this new value, and we haven’t upped the price at all ever since then. And so let’s do these organization features and then release them at a higher price point so that we can get people to pay more what we think is appropriate for the kind of features that they’re getting. There was this big jump between the personal plans, and the organizational plans, something like $75. You had to jump from like a $25 to $100 or something. We got a lot of push back on that, because people couldn’t stomach that much of an upgrade price. They wanted to go from 7, to 12, to 25, to something like 50. Instead we were like oh, you have to go all the way to 100. But you get all these features isn’t that great. And we talked about it a lot. What we ended up doing was realizing that our customer base is comprised primarily of small businesses and companies. Which makes sense because Git is a new technology, large companies hate risk, and so they’re not willing to make a big change in a [version troll] system overnight. But small companies, new companies they have no problem. They see something better and they use it. That’s what gives them the advantage of being small. After we had that realization, we said having a good upgrade path is drastically more important than trying to force people to pay what we wanted them to pay, or what we would have preferred them to pay. So we added another account in the middle, and split the difference and added that $50 per month account. Then people were extremely happy about that. So I think that was one thing that we realized was that really we were catering to thousands, and thousands of small businesses as much as we wish we had really big corporate clients on. We’re still comprised mostly of small businesses and so, having that realization allows us to really cater to them. It’s pointless to kid yourself about who your audience is. Cater best to your audience, and then you’ll get the kind of results that you want. This was about a year ago that we had this realization, and today there’s many more corporate clients but it’s still skewed very much towards the small. So we still have to think about that, and not kid ourselves about how huge we are, or how much penetration we’ve had into the corporate world. While it’s quite good now, it’s still far from where I think we want it to be. That was a pretty interesting thing to realize.