Decision review & postmortem routine
A practical playbook to routinely review important decisions, compare outcomes with expectations, capture learning, update decision criteria and playbooks, and assign follow-up actions so the organization improves decision quality over time.
Why this routine matters
Organizations that consistently learn from decisions become better calibrated, faster at making good trade-offs, and less likely to repeat avoidable mistakes. This playbook turns decision review from an occasional afterthought into a short, repeatable routine that preserves judgment, updates criteria, and drives practical changes.
Primary hunger
Learn from outcomes to improve future decision quality and calibration.
Cadence guidance
- Strategic decisions: review monthly or after the first meaningful evidence point (whichever comes first).
- Major operational bets or pilots: review quarterly or at pre-defined milestones.
- Small or routine decisions: bundle into periodic decision-scan sessions (quarterly) or use a lightweight checklist.
Who should attend
Decision owner, primary analysts, one or two stakeholders with opposing viewpoints (if applicable), and a facilitator/note-taker. Keep the group small enough to be candid and focused.
Preparation
- Bring the decision record (rationale, key assumptions, expected outcomes & metrics).
- Provide the most relevant outcome data and timeline (actuals vs expected).
- Share a short pre-read (1–2 pages) so the meeting focuses on learning, not re-presenting facts.
Agenda
- Revisit the decision record and assumptions
Confirm the original options considered, the rationale, and the assumptions about the environment, customers, costs, and timing.
- Compare actual results to expectations
Use the pre-shared metrics and data. Ask: which expected outcomes were met, which missed, and by how much? Were any external events material?
- Capture what changed and why
Identify which assumptions held, which failed, and whether the causal story fits the data. Distinguish between predictable variation and true surprises.
- Extract practical learnings
Turn observations into short, actionable lessons (for example: revised decision criteria, new guardrails, or triggers for scaling or sunsetting).
- Decide next steps
Choose whether to scale, iterate, pause, or sunset the initiative; assign owners and clear timelines for follow-up actions.
Decision record template (compact)
- Decision ID / Title
- Date
- Owner
- Context & problem — short description of why the decision was needed
- Options considered — including trade-offs
- Chosen action & rationale — explicit assumptions and estimated probabilities where possible
- Expected outcomes & metrics — what success looks like, measurement cadence
- Time horizon & review dates
- Risks & mitigations
Concrete prompts to surface useful learning
- Which of our assumptions were most important to the outcome?
- Were any important signals ignored or misinterpreted? Why?
- Which decision criteria need tightening, relaxing, or replacing?
- What would we change if we had to decide again today?
- How well were our probability estimates and risk assessments calibrated?
How to capture and apply learning
- Write short, specific lessons (one-sentence change + why) and tag them to the decision record.
- Update relevant playbooks, checklists, or decision criteria within the next sprint or review cycle.
- Record calibration notes (who estimated what, how accurate were estimates) to improve future probability judgments.
- Store both the decision record and the review outcome in a shared knowledge space with searchable tags (project, decision type, date, owner).
Follow-up and accountability
- Assign owners for each learning: who will update the playbook, who will adjust metrics, and who will implement guardrails.
- Set clear due dates and attach follow-up items to the owner’s regular planning or review process.
- Close the loop: confirm completion in the next cadence or via an automated checklist entry.
Facilitation tips
- Focus on evidence and causal reasoning rather than blame. Create psychological safety for honest diagnosis.
- Keep reviews time-boxed (30–90 minutes depending on scope). Aim for clarity and a small number of high-value actions.
- Use data visualizations to make gaps and trends obvious. Bring a neutral facilitator when stakes or politics are high.
Common pitfalls to avoid
- Confusing outcomes caused by external shocks with decision failures.
- Keeping learning in people’s heads instead of updating shared artifacts.
- Letting reviews become blame sessions or endless post-hoc rationalizations.
When to run a lighter review versus a full postmortem
- Run a lighter review for decisions that are routine or low-impact; capture one or two short lessons and move on.
- Run a full postmortem after unexpectedly large negative outcomes, major strategic pivots, or when repeated similar failures appear.
Starter checklist (use or adapt)
- Decision record prepared and shared ahead of time.
- Data and metrics accessible during the review.
- Owner and facilitator assigned.
- Three concrete learnings captured and assigned.
- Playbook or checklist updates scheduled.
Where this playbook can grow
This routine is a good candidate for a reusable decision-review toolkit that includes templates, interactive forms to capture reviews, automated reminders, and dashboards of decision outcomes. Teams can copy and tailor the toolkit to their context and scale it across the organization.
Discussion
Comments and conversation will live here.