Pre-mortem experiment that surfaces hidden assumptions in a hypothesis. Simulates failure, identifies testable claims across risk dimensions, and produces an experiment report.
Use when
Testable claims — each with a risk type and risk level — that must hold for the hypothesis to succeed. These become the input for data mining and customer validation experiments.
A structured risk profile: total assumption count by risk type, mandatory assumptions highlighted, overall risk assessment, and recommended next steps.
Project six months into the future and assume the hypothesis was proven wrong. The idea shipped but did not move the needle. Work backwards from that premise.
For each of Cagan's four risk dimensions — value, usability, feasibility, and viability — ask what failure would have looked like in that category.
Convert each failure mode into a positive, specific, testable claim that must be true for the hypothesis to succeed. Aim for ten to twenty assumptions.
Assess each assumption as mandatory, high, medium, or low risk. Record the full risk profile and highlight mandatory assumptions — these are the first to validate.
Retrospectives look backwards at what went wrong. Pre-mortems look forwards — imagining failure before it happens so the team can address risks while there is still time to change course. The technique forces you to bypass optimism bias by accepting failure as a premise, then reasoning backwards: if this failed, what would have been false? Each failure mode becomes a testable assumption that can be validated, monitored, or designed around before a single line of code is written.
Run before committing to significant discovery or delivery investment. The output feeds directly into data-mining experiments — the mandatory assumptions from the pre-mortem are the ones to test first.
Works better with