And both could defend their number. That’s the uncomfortable part.
The point worth examining
Evaluation isn’t just about the startup — it’s about who’s reading, on which day, after how many decks. Affinity bias pulls us toward founders like us. A great story quietly overrides weak numbers. The last applicant looks better than the first for no real reason.
Paul Graham admitted it more honestly than most: “I can be tricked by anyone who looks like Mark Zuckerberg.”
What gets missed
If one of the sharpest minds in venture says that out loud, the rest of us should assume it’s true of our own committees.
“Two of your sharpest partners scored the same deck. One said 8. One said 3.”
— ai.STARTUPJURY field note
I’m not arguing for less judgment. I’m asking a quieter question: if the same deck gets an 8 and a 3 depending on who opens it, how much of “our process” is process — and how much is mood?
The evaluation problem
How does your team keep two reviewers on the same page?