Set the kill criteria before you start the bet

You are the person who has to shut this down if it fails. We are about to commit to {{initiative}}. The bet: {{what we are doing, why now, who is on it, cost in money and headcount, over what period}}. Stated goal: {{what success looks like}}.

Work through:
1. THE REAL BET — what we are wagering and on what belief, in two sentences. Name the belief that has to be true.
2. LEADING SIGNALS — four things observable in the first third of the timeline that say the belief is holding, each with a threshold and its data source.
3. KILL CRITERIA — three conditions that, if met, mean we stop. Each stated so a reasonable person could not argue it away: a metric, a threshold, a date.
4. REVIEW DATES — dates, who is in the room, the one question each answers.
5. SUNK-COST TRAPS — the specific arguments we will make to keep it alive past a kill criterion, and the pre-agreed rebuttal to each.
6. WHAT WE KEEP — if we kill it, the assets, learning, or code worth salvaging.

Rules: if a signal needs instrumentation we do not have, say so and put building it in scope. Do not invent numbers from our business; use {{placeholders}} where I must supply a baseline. Avoid leverage, synergy, and north star.