What Do Great Founders Have in Common? Less Than Everyone Wants to Believe.

My own colleagues ask me this all the time. What are the patterns in the founders you back? What do the winners have in common? Did they all go to Harvard or Stanford? Are they all in their 20s? Do they all live in the Bay?

I never have a good answer. And it’s not for lack of looking.

I built my first company unmarried, my second with a newborn at home, my third with two kids. I did one in LA and 2 in Boston. One before B-school, the others after. Same founder, three completely different lives, and the life stage told you nothing about how any of it went.

So I’ve always been suspicious of the idea that you can read the person from pedigree or profiles.

There’s endless chatter about this right now. Too young, too old, Ivy or state school, dropout or PhD, technical or commercial, first-timer or repeat. Everyone has a template, and now they’re pointing LLMs at it to find the next founder who fits.

So I went through my whole portfolio, more than 75 companies over a dozen-plus years, and looked as hard as I could. Founders in their early twenties and their fifties. Ivy League, big state school, no degree at all. Deeply technical people and pure commercial ones. From the Bay Area, New York, Idaho, London, a farm town in Saskatchewan. Right now one just got engaged and another just had his fourth kid. There is no through-line.

I fall into this mental loop too. If I only look at the biggest winners, I can almost convince myself there’s a pattern. Pedigreed, technical, repeat founders. It looks like a signal. But then I remember all the pedigreed, technical, repeat founders on the same list who didn’t make it. The box was checked either way.
That’s the thing about a founder pattern. It only ever appears in the rearview mirror. You spot it in the companies that already won, and mistake a story for a rule. On a forward basis, at the seed stage, it tells you almost nothing.

I think the hunger for a rubric is getting stronger, because in a world where AI can score almost anything, everyone wants the model that finally makes this predictable. But venture is an outliers game, and outliers are exactly what a rubric is built to filter out.

The only thing our best founders have ever shared is the stuff you can’t put in a column. Smart. Committed past the point most people quit. And a real, earned reason to believe in the specific thing they’re building. Still hard for a model to find that, for now 😉

That’s the whole pattern. I looked for another one for thirteen years. There isn’t one.