How verdicts work, and why kitty says no

The process

Nine boards, in three stages. Seven research boards run in parallel — market, competitor, finance, legal, GTM, marketing, risk — each with live web search and each required to cite material claims and to mark estimates as estimates. The advisor board then reconciles the seven, including where they disagree. Finally the executive board reads everything and writes the verdict.

Where the score comes from

The opportunity score (0–100) is assigned by the executive board, working from the assembled evidence. It is not a weighted average of board scores, and there is no formula to reverse-engineer. The board is explicitly instructed to keep score and verdict consistent with the evidence — not to rate a brutal red-ocean idea a 90, and not to rate a strong one a 40.

This is worth knowing because it tells you how to argue with a score you think is wrong: point at evidence, not at arithmetic. Re-running a single board with better inputs will move the verdict; nothing else will.

The four verdicts

• GO — the evidence supports building it
• PIVOT — there's a real market here, but not in the shape you proposed
• CAUTION — viable under conditions the report names explicitly
• NO-GO — the evidence says don't

Why a NO-GO is the point

The stated enemy of this product is the six months you spend building something the market had already answered. A killed idea is not a failure, it is a life handed back. Ledger — the finance board — kills more ideas than the other eight combined, and he is not sorry.

This is also why the guarantee is phrased the way it is: kill one bad idea or you don't pay. See refunds.

Publishing a verdict

The Share button in the venture header publishes the verdict and copies the link. Sharing is entirely opt-in and off by default.

Published NO-GOs appear in the Killed Club, whose membership is earned one dead idea at a time. Published verdicts of any kind appear in the verdict library. Nothing is published without you clicking that button.