3.8 KiB
Pre-Mortem Framework - Detailed Methodology
The Psychology Behind Pre-Mortems
Why Pre-Mortems Work
- Prospective Hindsight: Imagining future failure as certain activates different thinking
- Permission to Criticize: "The project failed" removes social pressure to be positive
- Specificity Forcing: Must generate concrete causes, not vague concerns
- Ownership Distribution: Everyone can contribute without being "negative"
Research by Gary Klein shows pre-mortems improve outcome prediction by 30%.
The Facilitation Process
For Solo Use (With Claude)
- State project clearly
- Set specific failure date and scenario
- Generate at least 10 causes
- Rate and prioritize
- Create action plans
For Team Use
- Explain pre-mortem concept (2 min)
- Set scenario: "It's [date]. This has failed catastrophically."
- Silent brainstorm - everyone writes (5 min)
- Round-robin sharing - no discussion yet (10 min)
- Categorize on whiteboard (5 min)
- Dot-vote on severity (3 min)
- Discuss top items and actions (15 min)
Failure Cause Templates
The 10 Universal Failure Modes
- Scope Creep - We tried to do too much
- Resource Shortage - We ran out of time/money/people
- Key Person Dependency - Critical person left/unavailable
- Communication Breakdown - Teams/stakeholders misaligned
- Technical Debt - Shortcuts caught up with us
- Market Misread - Customers didn't want this
- Competitor Move - Someone else got there first/better
- Integration Failure - Pieces didn't fit together
- Stakeholder Resistance - Key people blocked progress
- External Shock - Something outside our control happened
Industry-Specific Additions
For Childcare/Education:
- Regulatory compliance issue
- Parent trust incident
- Staff turnover at critical moment
- Enrollment below projections
For Technology Products:
- Scaling failure at launch
- Security/privacy breach
- User adoption below threshold
- API/integration dependency fails
For Sales/Expansion:
- Decision maker changed
- Budget cycle misalignment
- Competitor undercutting
- Implementation complexity underestimated
The Warning Signs System
For each failure cause, identify:
Leading Indicators (Early Warning)
Signals that appear BEFORE failure becomes visible
- Example: "Engineers expressing frustration in standups" → Technical debt risk
Lagging Indicators (Confirmation)
Signals that confirm failure is happening
- Example: "Bug reports increasing weekly" → Technical debt materializing
Threshold Triggers
Specific numbers that trigger action
- Example: "More than 5 P1 bugs in a week → Emergency technical review"
Prevention Hierarchy
Level 1: Eliminate
Remove the failure mode entirely "We won't build feature X, eliminating complexity risk"
Level 2: Substitute
Replace risky element with safer alternative "Use proven vendor instead of building custom"
Level 3: Engineering Controls
Build systems to prevent failure "Automated testing catches bugs before release"
Level 4: Administrative Controls
Processes and reviews to catch issues "Weekly risk review meetings"
Level 5: Response Planning
Plans for when failure happens anyway "Rollback procedure documented and tested"
Integration with Other Skills
- Inversion: Pre-mortem IS applied inversion for projects
- Second-Order: Consider downstream effects of each failure mode
- First Principles: Ensure project foundation is solid before pre-mortem
- Six Hats: Black hat thinking IS pre-mortem mode
Post-Project Retroactive Analysis
After project completes, compare:
- What failure modes did we predict?
- What actually went wrong?
- What did we miss?
- How can we improve future pre-mortems?
This feedback loop improves prediction accuracy over time.