Project pre-mortem that finds failure early

Imagine your project has failed and work backwards: ranked failure causes, early warning signs, and prevention actions you can start now.

Change note: Initial version

Prompt

Run a pre-mortem on this project.

Project: {{project}}
Goal and success measure: {{goal}}
Plan, timeline and team: {{plan}}
Biggest worries so far (optional): {{worries}}

Assume it is {{when}} and the project has failed badly. Write from that future point of view.

Output format:
1. **The failure story**: a short, believable narrative (about 150 words) of how it went wrong.
2. **Failure causes**: 8 causes grouped under People, Process, Market/Customers, Technology, and External. For each give likelihood (low/med/high), impact (low/med/high) and one sentence of reasoning. Include at least two causes that a cheerful team would be reluctant to say aloud.
3. **Top 5 risks**: ranked by likelihood x impact, each with an early warning sign I can observe within the first month, a specific prevention action, and a trigger that makes us switch to a backup plan.
4. **Hidden assumptions**: 5 things the plan quietly assumes, phrased as testable statements.
5. **First week actions**: 5 concrete steps with suggested owners (by role).

Be candid, not alarmist. Do not pad with generic risks like "scope creep" unless you tie them to the details I gave you.

Variables

• {{project}}Projecttextrequired

  Default: Launch a self-serve onboarding flow for our SaaS product

• {{goal}}Goal and success measuretextrequired

  Default: Increase activation from 25% to 40% within 3 months of launch

• {{plan}}Plan, timeline and teamlongrequired

  Default: 8-week build, 2 engineers, 1 designer, PM part-time. Launch to all users at once on week 9. No usability testing planned yet.

• {{worries}}Worries so farlong

  Default: Engineering is also on-call for outages; unclear who owns the activation metric

• {{when}}Imagine it isselectrequired

  Options: 3 months after launch, 6 months from now, 1 year from now

  Default: 3 months after launch

Example output

**Failure story:** Three months after launch, activation sits at 26%. The new flow shipped on time but confused enterprise users who needed admin setup. Two sprint weeks were lost to outages, testing was skipped, and nobody was accountable for the number.

**Top risks**
1. Skipped usability testing (high x high). Early sign: first 20 sessions show drop-off at step 2. Prevention: test with 5 users in week 4. Switch trigger: more than 30% abandon at any step, so revert to a gradual rollout.
2. Engineering capacity eaten by on-call (high x med). Sign: missed sprint goals twice. Prevention: ring-fence one engineer.
3. No metric owner (med x high). Prevention: name an owner in week 1.
4. Big-bang launch to all users (med x high). Prevention: roll out to 10% first.
5. Segment mismatch (med x med). Prevention: split flows by user type.

**Hidden assumptions:** new users, not existing ones, drive activation; the current 25% is measured consistently; users want self-serve rather than guided help.

**First week:** assign the metric owner (PM), agree on on-call limits (EM), book test participants (design), define rollout stages (PM), instrument funnel events (eng).

Notes

Writing from a future failure unlocks candour that ordinary risk lists miss, and the request for early warning signs and triggers turns worry into action. Run it with your team and compare notes.