Core Idea
Definition
A postmortem is a structured retrospective that compares expectations, actual outcomes, causes, and lessons in order to improve future decisions and execution.
In Plain English
After something happens, ask what we expected, what really happened, why the difference occurred, and what we should change next time.
Framework Structure
Components
Flow
Reconstruct expectations -> Compare with reality -> Analyze why -> Extract changes worth keeping
How to Apply
- 1.Capture what the team expected before hindsight rewrites the story
- 2.Describe what actually happened as concretely as possible
- 3.Separate luck, execution, assumptions, and environmental change
- 4.Identify which lessons are actionable and generalizable
- 5.Turn the best lessons into process updates, not just observations
When to Use
- •After launches, incidents, projects, and negotiations
- •When outcomes differ from expectations
- •When a team wants to learn without defaulting to blame
- •After both failures and surprising successes
- •Any context where repeated decisions can be improved through structured review
When NOT to Use
- •When the review is being used mainly to assign guilt
- •When there is no reliable reconstruction of what was believed beforehand
- •When the environment changed so much that no meaningful lesson transfers
- •When the team will not act on the insights
Example
Problem
A product launch underperforms despite internal confidence.
Application
- 1.Review what the team predicted before launch
- 2.Document actual adoption, support load, and user feedback
- 3.Separate causes such as flawed assumptions, execution gaps, and plain bad luck
- 4.Translate the best lessons into future launch checks and research practices
Conclusion
The review improves future decision quality because the team learns what to change instead of merely reliving disappointment.
Takeaway
A good postmortem turns outcomes into better future judgment by preserving honesty about both uncertainty and process.
Common Mistakes
- •Letting hindsight rewrite what was knowable at the time
- •Confusing bad outcome with bad decision
- •Extracting lessons that are too vague to use
- •Overfitting one event into a universal rule
- •Reviewing only failures and ignoring lucky wins or hidden near misses
How to Practice
prediction capture
Write expectations down before results arrive so the later review has a reliable baseline.
process vs luck split
For each result, ask separately what was due to reasoning, execution, and chance.
lesson to change
Turn each major lesson into a concrete behavioral or process adjustment.
Related Cognitive Biases
hindsight bias
People tend to reconstruct the past as more obvious than it really was.
outcome bias
Bad luck can make good decisions look bad unless process and result are separated.
self serving bias
Teams may credit success to skill and blame failure on circumstance unless the review is disciplined.
Related Frameworks
Related Skills
Variants & Extensions
Typical Failure Modes
- •Blame theater
- •Hindsight distortion
- •No implemented learning
Further Reading
- Black Box Thinking by Matthew Syed
- Thinking in Bets by Annie Duke
- The Checklist Manifesto by Atul Gawande