Four and a half years ago, I decided to write a Mac tool for Git and Mercurial, which I’d eventually name SourceTree (aside: names are hard, and I was quite pleased with this one). I wasn’t happy with the Mac apps that were out there at the time and thought I could write something that fit my needs better, and by extension the needs of other developers who felt like I did. I’d never written a Mac-specific app before, and I thought it would be fun to learn how. I also knew that the Mac was a platform where independent app developers could make a living on ‘premium’ apps, so it made business sense. I could see the beginnings of the trend away from Subversion to git/hg and felt there would be an expanding market for such a tool. So, I took a punt.
With a little apprehension I quit all my contracting gigs, bought a couple of books on Objective-C and Cocoa and got my head down, trying to ignore the dwindling bank balance as the months passed. The words JFDI, Lean Startup, Customer Development all rang in my ears. 6 months later, and almost exactly 4 years ago today, I announced the launch of SourceTree version 1.0 on this blog. It was a slow start, but I was prepared for that, and in many ways it was a boon because it gave me time to refine the product with a small but growing gang of early adopters. I took a little more contracting work on the side for a few months to stabilise my finances while I iterated, but only 5 months after that initial release, SourceTree passed an inflection point and adoption started climbing – it was in profit and growing at a good speed. Around this time Atlassian approached me about an acquisition; at the time I was happy to just have a profitable product that meant I could afford to work on what I wanted, but they were convinced that together we could take this product to a whole other level. I eventually decided to accept, and boy, were they right.
The blur of the following 3 years included: porting SourceTree to Windows; hiring fellow Guernsey developer Kieran Senior whom I mentored with all the subtlety of the Eye of Sauron; morphing into a team spread across Guernsey/Amsterdam/San Francisco/Sydney with a real designer, support and marketing folks, and above all, watching a ton more people start to use SourceTree. More than I could ever have imagined. When I stepped down from Ogre I wondered whether I’d ever build anything that popular again, but it turns out the answer was yes, and then some. It’s been a pleasure, and my thanks to everyone who helped make this possible.
However, as much as I love SourceTree and have enjoyed this amazing journey, it’s time for me to seek a new challenge. Four and a half years is a long time to work on one product, and there are so many other problems I’d like to solve. We hired Mike Minns earlier this year to lead the team, he’s a seasoned lead developer located in Sydney (and ironically an ex-pat – we don’t deliberately recruit everyone from the British Isles, honest!), and together with Kieran will be responsible for all future SourceTree development, while I bow out and take a merely advisory role when needed – which will likely comprise of ‘Steve, what the f*** is this code??’ ;).
I’m starting work on one of the (many) projects I shelved a while ago – unsurprisingly it’s one of those problem areas that bugs me personally (what else?) and I’m pretty certain that other people will see the merit in addressing it the way I have in mind too. Hopefully I’m still with Atlassian, because they’ve been great about giving me freedom and this particular project very much suits doing within this environment. I’ll talk more about this when the time is right of course. *tease*
I’m posting about this on my blog because a number of people know me personally as the go-to guy for SourceTree, and sometimes short-circuit the official reporting channels for feature requests and bugs on Twitter and email. This was fine, I always like to connect personally with people using my products, but obviously in the future this won’t work any more. Thanks for all the feedback and support over the over the last 4 years, I’m so lucky to have managed to make another tool that developers really connected with. SourceTree will continue without me and I’m sure we’ll see great things from the team in future. Now I’m off to see if I can do something else useful