how i read your startup in the first 30 minutes

i’ve done thousands of these calls. somewhere along the way it stopped being evaluation and became pattern matching. within 30 to 60 minutes i can usually see where you’re going to fail. sometimes i tell you. sometimes i don’t, because people don’t like hearing they’re wrong from a human. so here’s the whole thing written down. read it before you message me and you’ll save us both an hour.

what i ask first what are you building, in your own words. how much are you raising and at what valuation. if you don’t know the second one, we’re early, and that’s fine, but say so.

what i look at before you say anything your deltav profile, if you have one. your twitter. your dune or google analytics, if connected. whether your numbers on the profile match the numbers you’re about to tell me. stale numbers on a profile are a tell.

the questions, in the order i ask them

who pays, and is it recurring. per-report is a dev shop. free users are not users.
what’s the number you look at every week, and what did it do last month. if you say a round number, i know you read it off a slide.
out of everyone who signed up, how many came back last week. 500 users and 100 returning is interesting. 250,000 users and no returners is not.
who are your first 50 customers, by name or by type. “creators and b2b” means no one.
what’s your closest competitor, and what’s the delta. 30 percent better means nobody switches.
are these contracts, lois, or soft commits. i’ll ask this about every logo on your partnership slide.
is your previous investor doubling down. if not, why not.
is this a consensus market or a contrarian one. if consensus, where’s the speed. if contrarian, where’s the proof it works.
what will you not leave when it gets hard. if the answer is “this, obviously”, i’ll ask again differently.
what have you said no to this month.

what counts retention. a chart with a bend in it. a paying customer who’s come back. a previous investor writing a second check. a founder who’s lived the problem. a narrow icp with a number in it. a founder who’s already pushing on something.

what doesn’t ideas, unless you’ve exited before or run engineering at a coinbase. mvps, since anyone can build one in a week now. patents. advisors. mous. social impact. a program you got into. the tam slide. being on a chain. “we’re first.”

what i’ll do if it’s there push you to 100 users this week and ask again friday. introduce you to two people, not twenty. put you in front of investors only once i’m sure you’ll close, because a demo day where nobody raises is a waste of everyone’s time, including yours.

what i’ll do if it’s not tell you, if i think you’ll listen. send you a book or a video instead of tactics, because the problem is usually three levels under the one you asked about. and tell you to come back when you have retention, because that’s the only thing that’s ever worked.

the fifteen places founders fail, in rough order of how often i see them coding instead of selling. an icp that’s everyone. mistaking free users for traction. waiting for a retweet to be a go-to-market. growth without retention. a pitch that explains the problem instead of the curve. a raise that bundles the whole vision. announcing before the wire. believing a vc’s nice words. taking money that brings no distribution. pivoting with a smooth story and a sunk cost. treating investors like parents instead of crew. launching a token before the numbers. moving in a comfortable room. and the one under all of them: not having decided what they won’t quit.