Frame pre-mortems as explanations of failure, not predictions
Frame pre-mortem prompts as 'It is [future date]. This has failed completely. Write why.' rather than 'What could go wrong?' to shift cognition from speculation to explanation.
Why This Is a Rule
"What could go wrong?" triggers predictive reasoning — your brain generates a short list of obvious risks, the same ones it always generates, and stops. The question feels answered because you can picture a few failure modes. But predictive reasoning is lazy by design: it satisfies the question with minimal effort.
"It is [future date]. This has failed completely. Write why." triggers explanatory reasoning. The failure is assumed. Your brain shifts from "could this happen?" (usually answered "probably not") to "why did this happen?" (which demands specific causal chains). Research on prospective hindsight shows this framing increases the number of identified risks by 30% and produces qualitatively different risks — systemic, second-order, and organizational failures that predictive reasoning misses entirely.
The difference is grammatical but cognitive: prediction asks if, explanation asks why. Explanation is harder and produces more.
When This Fires
- Project kickoff: before committing team resources to a plan
- Product launch: before shipping to users
- Hiring decisions: before extending an offer
- Any commitment where failure would be costly and reversing would be difficult
Common Failure Mode
Running a pre-mortem with "what could go wrong?" framing and feeling satisfied with the result. The team generates five obvious risks, assigns mitigations, and moves on. The risks that actually kill the project — the ones involving coordination failures, incentive misalignment, and unstated assumptions — never surface because predictive reasoning doesn't reach them.
The Protocol
Write the exact prompt: "It is [6 months from now]. [Project name] has failed completely. The team is doing a post-mortem. Write the explanation of why it failed." Give everyone 5 minutes to write independently before sharing. The independence prevents anchoring — if the first person says "scope creep," everyone else's explanation shifts toward scope creep. Collect all explanations, then categorize and prioritize.