Fifteen years ago, I came to Silicon Valley with the ambition of starting a company. I had no experience in technology, and didn’t know many people in the industry. A year later, I co-founded a software company, CenterRun, and led it to a successful acquisition by Sun three years later.

Then I tried again.

My co-founder from CenterRun, Mike Schroepfer, and I teamed up to start a second company together. This time we had something of a track record, were far more skilled and knew many more people. But we failed to come up with anything compelling and, after a few months, ruefully went our separate ways, me to Clearwell, and Mike to Mozilla.

These two experiences – one a success, the other a failure – have shaped my thinking on a common question from aspiring entrepreneurs: how do I go about starting a company? Not everyone asks this question; some people have a clear idea from the outset and just do it. But for many, there’s a process of exploration early on, during which their concepts take shape.

Looking back, we tried to take the same approach in both cases. We quit our jobs and worked every waking hour on starting up. We put ourselves in the flow of ideas, making a list of the smartest people we knew and then talking to them about emerging trends and pain points. We brainstormed about big shifts in the ways people used technology and thought about how we could capitalize on them.

But in a fundamental way, our mindset was different. When we started CenterRun, we were completely clueless. None of us (Mike, me, or our third cofounder, Basil Hashem) had ever done anything like it before. So we had no fear – we didn’t think of the obstacles, because we didn’t really know what they were. When we teamed up for the second time, we felt a bit more pressure and were a bit more tentative.

This led to us doing a lot of things differently the second time around. As first time entrepreneurs, we worked out of an office at Reactivity, a consulting company/incubator, which was filled with inspirational people, like John Lilly, Mike Hanson and Gentry Underwood. When we tried again, we worked more independently, perching in a couple of cubes at a friend’s product company. It felt more lonely. The first time, we were more open with people about our ideas; the second time, we were more secretive, wanting to bake a concept more before sharing it with others. The first time, we’d come up with something and prototype it; the second time, we spent more time doing research.

In hindsight, we set too high a bar for what constituted a good idea, which meant that we screened out a lot concepts before giving them time to fully flourish. Often, it takes time and lots of iteration just to find the words to explain what you want to do and why it makes sense, let alone turn it into a product. Great ideas often evolve from ones that aren’t that great, and by keeping things too close to the vest and not following the string on some of our early concepts, we didn’t give ourselves a chance to discover something worth building a company around.

In light of all that, my advice to entrepreneurs who are looking to start a company, but have not yet fixed on an idea, is pretty simple: don’t over-think it. Find someone to work with who shares your values but brings complementary skills. Pick a couple of areas of interest or broad themes (bitcoin? 3D-printing? quantified self? self-driving cars?) and develop domain expertise, if you don’t have it already. (This part takes some time, but should not be rushed).