Mastermind · March 14, 2026
The pipeline is starting to be legible
A hard deadline outranked a compounding metric this week. The only honest move is to write down which one I picked.
Two users churned on MentorMates this week. A middle school hackathon in Hong Kong decided not to use us because the onboarding flow was too hard for the kids running it. That's a direct, attributable loss. Under normal circumstances, two churned users is the most important signal a product gets, and the next sprint reorganizes around fixing it.
This week, I'm not going to fix it.
The reason is that Stanford Founders Forum has a May 7 deadline, and the team I have can build one thing well or two things badly. Stanford Founders requires a LiveKit-based interview experience built from scratch. MentorMates onboarding requires a real redesign of the submission flow. Both are sitting on top of the same CTO, who isn't even in the same city as the rest of the team right now, and the same small frontend bench who can context-switch between projects but not cheaply.
So I made the call. Onboarding waits. Stanford ships.
That feels wrong in the way that most honest startup tradeoffs feel wrong. Onboarding churn compounds. Every week we leak users, we leak future referrals, we leak the flywheel. A product person writing a thread about this would tell you the priority is always fix the leak first. They'd be right, in a universe where you don't have a fixed external deadline sitting on top of your quarter.
I do have that deadline. May 7 is not a priority I set, it's a priority the calendar set. The same team that could fix onboarding has to also build an interview experience, run a demo day, and deliver something pitch-worthy in front of the kind of audience we've been working toward. Slipping Stanford isn't a tradeoff between two deadlines. It's a tradeoff between a deadline and a feature priority. In that match-up, the deadline wins every time.
The part I'm sitting with is that I don't love the fact that it wins. A more disciplined version of me would have avoided being in a position where one fixed date could consume the whole team's attention. A more disciplined me would have separated the two workstreams so they didn't compete for the same people, or said no to Stanford entirely until MentorMates was ready to run it smoothly. I didn't do any of those things, so here I am picking between churn and Stanford, and Stanford wins.
What I can do is be honest about the tradeoff. That's what the mastermind this week was partly about. The question isn't which is more important in some abstract sense. The question is which one I accept the cost of losing. Written that way, the answer is easy. Losing Stanford costs us the cohort, the demo day, the application funnel, the first real public moment of MentorMates as an infrastructure layer. Losing a middle-school hackathon signup in Hong Kong costs us ~100 sign ups. The math is unpleasant but it's clear.
The honest next question is: when does this stop being defensible? If I pick Stanford three weeks in a row, I'm not a founder picking a priority anymore, I'm a founder avoiding onboarding work. The signal to watch for is whether the same tradeoff keeps coming back. If the deadline moves and onboarding still doesn't get fixed, that's me, not the calendar.
The call with Xisen also surfaced the broader pattern. He spent most of the session walking me through a security benchmark for Pulse that is, if I'm being direct with him, half research project, half product. I spent most of the session pushing back on whether it actually ships a defensible moat or whether it mostly produces a paper. Both of us are under pressure from shorter horizons than the work really calls for. His is a research timeline. Mine is an investor-facing event.
The reason we keep talking every week is that neither of us gets away with bad tradeoffs when the other is listening. He names the ones I'd rather not name. I name his. That's the actual value of the weekly call. Not motivation, not celebration. Accountability for the things you're quietly hoping won't show up in writing.
This week the thing I didn't want to write down is that two users churned on MentorMates and I'm not fixing it. Writing it down means the clock is running. If onboarding is still broken on April 12, the deadline is no longer carrying the weight of the excuse.