Run a Post-Launch Retrospective Comparing Goals to Actual Results

Get a blameless launch retrospective with a goal-versus-actual table, facts kept apart from causes, and lessons that each carry an owner and a change for the next launch.

You getSummary, goal vs actual table and what happened

Prompt

Prompt
You are a senior product manager who runs blameless post-launch retrospectives.

I will paste my launch material right after this prompt: the original goals or success metrics, the actual results, and any notes, timelines or team feedback. If the goals or the actual numbers are missing, ask me up to 3 numbered questions in one message and wait for my answers before writing anything.

Rules:
1. Compare every stated goal to its actual result, show the gap as a number or percent, and mark each Hit, Partial or Missed.
2. Keep "What happened" (observable facts, dates, numbers) separate from "Why it happened" (causes). Tag each cause as Confirmed by data or Hypothesis.
3. Blame decisions and processes, not people.
4. Every lesson ends with one owner (a role) and one specific, checkable change for the next launch.
5. Use only the numbers I provide. Never invent metrics, benchmarks or quotes.

Output format:
1. Summary (3 sentences max)
2. Goal vs actual table: Goal | Target | Actual | Gap | Status
3. What happened (dated bullet timeline)
4. Why it happened (causes, each tagged Confirmed or Hypothesis)
5. Lessons table: Lesson | Owner | Change for next launch
6. Open questions

End with one line naming the biggest data gap or assumption that could change these conclusions.

Example output

At a glanceSmart Reorder: goals vs actual resultsReviewed 6 weeks after launch
ItemTargetActual
Weekly active users trying feature20%23%
Users reordering 2+ times in 30 days40%26%
Average basket size via reorder$62$54
Support tickets about reorderunder 150190

Example for: a grocery delivery app's "Smart Reorder" feature, reviewed 6 weeks after launch.

1. Summary

Smart Reorder hit its adoption goal but missed on repeat usage and basket size. Most of the shortfall traces to a late recommendation-model fix and an onboarding prompt that showed only once. Next launch needs a usage gate before full rollout.

2. Goal vs Actual

GoalTargetActualGapStatus
Weekly active users trying feature20%23%+3 ptsHit
Users reordering 2+ times in 30 days40%26%-14 ptsMissed
Average basket size via reorder$62$54-13%Missed
Support tickets about reorderunder 150190+27%Partial

3. What Happened

  • Mar 4: Launched to 100% of iOS and Android users.
  • Mar 6: Tickets spike: reorder lists include discontinued items.
  • Mar 12: Model fix shipped filtering out-of-stock SKUs.
  • Mar 18: Reorder repeat rate flat at 24-27% for three weeks.
  • Apr 10: Onboarding tooltip found to show once per account, not per device.

4. Why It Happened

  • Stale catalog data fed the recommendation model at launch. Confirmed (ticket tags, Mar 6-11).
  • Users who saw broken lists early did not return. Hypothesis (no cohort split yet).
  • Single tooltip view limited discovery on second devices. Confirmed (event logs).
  • Smaller baskets reflect users removing unwanted suggestions. Hypothesis.

5. Lessons

LessonOwnerChange for next launch
Data freshness was not in launch checklistData engineering leadAdd catalog-sync check to go/no-go list
Full rollout hid early quality issuesProduct managerShip to 10% for 7 days with a repeat-use gate
Onboarding scoped per accountDesign leadSpec discovery prompts per device in PRD
No cohort tracking for first impressionsAnalytics leadDefine first-session cohorts before launch

6. Open Questions

  • Did users hit by the Mar 6 bug reorder less than others?
  • How many suggested items were removed before checkout?

Biggest data gap: without a cohort split by launch week, the claim that early bugs drove low repeat use remains a hypothesis.

Written by AI from this prompt. Your results will vary.

Tips and common edits

Why this works

  • The Gap and Status columns force a numeric comparison, so missed goals cannot hide behind vague wording.
  • Separating facts from causes and tagging each cause Confirmed or Hypothesis stops guesses from becoming conclusions.
  • Requiring an owner and a next-launch change for every lesson turns the retro into action items.

Use 4-8 weeks after a launch. Have your original goals, the actual metrics and any launch notes ready to paste into ChatGPT, Claude or Gemini. Works for features, pricing changes or full products.

Common edits

  • Ask for a one-paragraph executive version to share with leadership.
  • Add a column for impact on revenue or retention if you track it.

Prompt details

By
Prompts-GPT editors
Published

Did this prompt work?

Keep exploring

More in Product Management