From “No Ideas” to a Great Idea

So in the background this whole time, I've been wanting to start a company. I'd worked at these really small, EasyBloom was the smallest, about 8 people. I worked at big companies. I'd launched business divisions, and I was like, well, this is great. I knew that I could go from zero to a million, but I want to do it for myself. I don't want to do it for somebody else. I don't want to do it as part of a spinout. I don't want to do it as part of a division. I want to do it for me. The issue was I had no ideas. I was like, I'm a very boring person. All I really know how to do is build software. That's it. I'm not somebody who's like, oh, I also know how to do crazy ideas all my friends were doing. I was just like, all I really know is software itself. I know how to ship a release. I know how to design a release. I know how to bring it to market. But just software stuff. That's what I love. I love software. I mean, my friend John Kodumal and I, who'd gone to college together, we would hang out. He was in engineering, and we would just sit around and talk about software. How should Agile be implemented? Silly stuff we'd seen product managers do. How to build better releases. Release management. That was all we talked about. So John and I kept trying to come up with an idea to start our company around. Finally, we decided that we would just start working on something 10 hours a week. So if there's ever anybody you're considering joining a company with, practice it first. You don't know what it's like until you actually work with somebody what it's like to work with somebody. Sounds obvious. So what we started doing is we started working for 10 hours a week. We both kept our full-time jobs. John was at Atlassian, I was at Concur. and we would work on our idea for 10 hours. And we cycled through about, depending on how you count it, like 3 to 7 ideas. Initially, we thought about doing an app for children. Apps for children are actually very hard because there's a lot of regulations around what content you should show children. We thought about doing a productization of his PhD thesis, which was around static analysis, and we actually worked on this for a long time. We worked on this for around a year. So I was out there talking to a lot of developers. It just didn't feel right. And then when we were talking to people about ideas, suddenly it became really clear to us that we should do something around feature flags. I was talking to my developer friends. They said, you know what I really like and miss is the framework that I used to have at Facebook. or the framework I used to have at Amazon, or I could see what people at Dropbox are building. I'd also, we'd had an internal framework at TripIt. We had this internal framework where we could roll features out to different users, we could control it, we could see who was using a feature well. And I had always complained when I was at TripIt, like, why isn't this a product? Why do we have to keep doing it in-house? Even when I was back at Monster.com, I remember we had this catastrophic release once where I really wanted to try two different versions of something and then be able to do a kill switch, and there just wasn't a way to do it. I was like, why don't we build a product that makes software development better? And John said, that's it. That's our idea. Finally, we have found an idea that takes advantage of what we're both really good at, which is software. Like the stuff that we thought wasn't important was actually really important because we understood software. John had been at Atlassian, he had a PhD in software, CS. I knew software inside out. So we started LaunchDarkly.