Notes on early-stage investing

At the earliest stages, there is almost nothing to analyze in the traditional sense. No revenue, no retention curves, sometimes no product. What you are really evaluating is a person and a thesis about how the world is going to change.

I look for founders who have earned their insight — who have lived inside the problem long enough to know something the market doesn't. Pattern-matching on pedigree is a lazy shortcut and it shows in portfolios.

The second thing I look for is velocity of learning. Early companies are hypotheses. The founders who win are the ones who cycle through hypotheses fastest without losing their conviction about the destination.

Finally, I ask myself whether I would enjoy a three-hour conversation with this person every month for a decade. Because that, roughly, is the commitment.