Reconstruct what was known before launch
Collect the original goals, assumptions, scope, timeline and decision records. Preserve the version that existed before results were visible. Rewriting the original objective after the outcome makes it difficult to learn from the choice.
Build a timeline of important events: readiness checks, changes, communications, incidents and customer responses. Use dated records where possible and label recollections as recollections.
Separate delivery from the underlying proposition
Ask whether the launch reached the intended audience, delivered the promised experience and produced the expected customer response. Failure at an earlier stage can prevent a fair test of a later assumption.
For example, an illustrative launch with broken checkout cannot establish that customers rejected the price. Record the technical failure first, then identify which demand questions remain unanswered. Avoid using one outcome to explain every part of the system.
Use outside cases as diagnostic prompts
Find cases with a similar failure mode or coordination challenge. Use them to generate questions about sequencing, ownership and safeguards. Do not import another company’s explanation as the cause of your own result.
Compare the evidence for competing explanations. Identify which finding is supported, which is plausible and which needs a follow-up test. Include successful parts of the launch so improvements do not accidentally remove working practices.
Turn lessons into bounded changes
Each action should have an owner, a completion condition and a place in the next launch process. “Communicate better” is too vague; a named readiness review with specific inputs is inspectable.
Review whether the changes actually occurred and whether they reduced the targeted failure. Keep the list short enough to execute. A postmortem is unfinished when its lessons exist only in a slide deck.
Use this worksheet for your decision
Copy this into your notes. Leave unknowns visible instead of filling gaps with guesses.
Original goal and assumptions: Pre-launch evidence: Timeline: Delivery failures: Untested demand questions: Competing explanations: Practices to keep: Action, owner and completion condition: Next-launch verification:Download the text worksheet ↓
Find evidence in the archives
Choose a supporting case and a counterexample with a different outcome. Read their sources and check whether their conditions match yours.
Cases to practice with (read and check the underlying sources):
Bring your question to the cases
Search a company or topic in the app, read the cases and ask follow-up questions to compare approaches.