Why I left Google

tree-small
Make like a tree and leave

When I decided to leave Google a few months ago, I was asked all the questions you might reasonably expect people to ask. “Isn’t it a fantastic place to work? Couldn’t you just switch projects to something really exciting like self-driving cars? Don’t they pay really well? Are you sure you still know how to fend for yourself after all of the free food? I’d cut off a leg if it meant getting a job at Google, what’s wrong with you?” And so on.

Since this blog is supposed to document my post-employment adventures, I guess it makes sense to start at the beginning: Why I left Google, and why I’m not planning to be an employee again for the foreseeable future.

Was it because I no longer liked the project I was working on? No. My last project was Google Chrome. It’s a pretty fantastic project, I’m a fan of the product, I was working with a lot of incredibly great people, and it was, for a long time, an incredible learning experience to be part of such a massive engineering undertaking (last month, over 600 user accounts made code changes), witness the problems you start having at very large scale, and participate in fixing them. To be honest though, I enjoyed myself quite a bit more when I worked on much smaller teams that owned an entire product end to end, such as the Google Desktop team that was never more than about 18 or so engineers and about 30 people in total, yet produced and supported a product used by tens of millions of users every week.

Was it because, for the last three years of my nine and a half years at Google, I was working remotely from Iceland? Not really. The Chrome project is a fantastic project for working remotely, because its culture fosters contribution from teams and individuals all over the world. I traveled a fair amount, enough to see the core team in California quite regularly, but not so much that it became a drag. If I’m being honest though, I think I might have stayed at Google for a bit longer if I’d been working from Google HQ in Mountain View, California. There is just so much Kool-Aid to drink there.

Was it because I felt like I was no longer learning new things and growing in my role at Google as fast as before? Maybe a little bit. For my last couple of years there, I started feeling that maybe Google wasn’t the place where I was going to do the best work of the rest of my life. But that wasn’t the only reason. If it had been, it would have been quite easy to let things slide, stay in the cushy job, maybe switch projects to something where I could learn a bit more.

I think there are two separate reasons I decided to leave. One is that I have unfinished business, and the other is that I want to be the master of my own destiny to the largest extent possible. Let’s start with the latter.

I went through two different project cancellations at Google, of projects that I was leading at the time of cancellation. One of them I led from its inception, the other I had taken over about a year prior. Both of these projects were innovative, risky, and were significantly affected by outside forces – competition and a shifting technology landscape – and were cancelled because Google didn’t want to invest more in the face of those outside forces.

I believe the best work of the rest of my life will also involve risky projects.

When you do a risky project at a big company, there will always be a decision maker somewhere above you with the authority to pull the plug, regardless of how much you are willing to fight for the project, make the case, and pledge your blood, sweat and tears to make things work. To make matters worse, when the plug is pulled, the project might completely disappear from the face of the earth.

When you do a risky project as an entrepreneur and things are not quite going your way, your investors might get spooked, you might have to pivot, pull all-nighters, scale down your company, fire some of your best friends, pull favors, piss people off, and generally do incredibly difficult things. However, your ability to figure out ways to continue the project if you still believe in it is much greater. Worst case, you can probably open-source some work that could prove useful to others.

There are other ways to get to finish risky projects, you might say. Go into academia, or join a research department somewhere. You might not get to build full products but you’ll generally be able to find ways to keep researching whatever you’re interested in. This is where my unfinished business comes in.

Before Google, my career in software was exclusively at startups. I was an early employee at Iceland’s first Internet startup, I co-founded a spin-off from that company, co-founded another company on the side that I ended up not joining full-time, and was an early employee at another couple of startups before joining Google. But I never had full responsibility for any of these companies, and I didn’t see any of them through to successful exit or profitability. That said, a few of them did enjoy significant success, ranging from moderate to Icelandic-entrepreneurship-landscape-changing.

Bottom line: I have the bug. I need to build my own company. I have a lot of useful experience that I can build on.

I might be crazy. The statistical track record of entrepreneurship says that I probably am.

But hell, here goes nothing.