How to Run This (60 minutes)

Prep (before meeting):

1. Everyone fills out their section
2. Keep it to one-liners with specific numbers/names

Meeting:

1. Run thru (30 min): Run through the notes and actionables
2. Assign (15 min): Each problem gets one DRI and deadline
3. Document (15 min): Share in Slack immediately

Rules:

1. One-liners only - if you can't say it in one line, you haven't clarified the issue
2. Use real numbers: 2,000 users, 4 weeks, $4M, 80% reduction
3. Name names in commends: "Alex did X" not "engineering helped"
4. Every problem needs YOUR contribution - no blaming external factors
5. Solutions must have deadlines

Retrospective

1. Person
   1. Wins
      1. Personal:
         1. Things you're proud of having achieved. One liners.
      2. Commends:
         1. Things that you appreciate and want to highlight from other people. One liners.
   2. Problems
      1. Issue: One liner
         1. What I did to help create this situation: one liner
         2. Proposed solution: one liner

Example: Q3 Fintech App Team

1. Sarah - Product Manager
   1. Wins
      1. Personal:
         1. Cut instant transfers scope by 70% and still hit $4M daily transaction volume
         2. Reduced PRD review cycle from 2 weeks to 3 days by killing committee reviews
      2. Commends:
         1. Alex built adapter for payment API change in 4 hours, saving our Q3 launch
         2. Jordan killed her own "beautiful" slider design when data showed 92% type amounts
   2. Problems
      1. Issue: Budget calculator crashed for 2,000 users with >50 linked accounts on launch day
         1. What I did to help create this situation: Only tested with my own 5 accounts; never checked actual user data
         2. Proposed solution: Pull user distribution data for every feature; add to PRD template by Oct 10
      2. Issue: Lost 15 engineering days to platform consistency debates across 3 features
         1. What I did to help create this situation: Didn't set iOS-first or Android-first decision upfront
         2. Proposed solution: Page 1 of every PRD states lead platform; other follows without debate
2. Alex - Senior Engineer
   1. Wins
      1. Personal:
         1. Shipped instant transfers in 4 days by reusing payment webhook handler from last quarter
         2. Reduced API response time from 3.2s to 0.8s by fixing N+1 query in account aggregation
      2. Commends:
         1. Sarah killed 8 features from instant transfers; cleanest PRD I've ever implemented
         2. Jordan provided Figma specs with exact padding values; zero design-dev ping pong
   2. Problems
      1. Issue: Budget calculator crashed for 2,000 users with >50 linked accounts on launch day
         1. What I did to help create this situation: Loaded all accounts into memory; never tested beyond 10 concurrent users
         2. Proposed solution: Thursday 3pm weekly load test with 10x production traffic before any launch
      2. Issue: Search feature slipped 4 weeks after discovering Elasticsearch version incompatibility
         1. What I did to help create this situation: Said "fuzzy search is easy" without checking our ES version supports it
         2. Proposed solution: 2-hour technical spike required before any feature commitment; results in PRD
3. Jordan - Designer
   1. Wins
      1. Personal:
         1. Instant transfers built in 4 days using only system components; zero custom UI
         2. Reduced design QA bugs by 80% by adding interaction videos to every Figma frame
      2. Commends:
         1. Alex implemented complex transfer animation perfectly after one video walkthrough
         2. Sarah pulled support tickets proving users don't want "delightful" features; just working ones
   2. Problems
      1. Issue: Budget calculator crashed for 2,000 users with >50 linked accounts on launch day
         1. What I did to help create this situation: Designed cards layout for max 5 accounts; never made list view fallback
         2. Proposed solution: Every design includes stress test: 0, 1, 10, 100+ items documented in Figma
      2. Issue: Wasted 2 weeks on high-fidelity designs for technically impossible features
         1. What I did to help create this situation: Went straight to pixel-perfect without engineering input
         2. Proposed solution: Wireframes only until Alex confirms "buildable in sprint"; then high-fidelity

---

Team Problem Inventory

After individual sections, vote on top 3 to fix in Q4:

1. ✅ Budget calculator crashes → DRI: Alex → Fix by: Oct 15
2. ✅ Platform debate waste → DRI: Sarah → Fix by: Tomorrow
3. ✅ Design before feasibility → DRI: Jordan → Fix by: Monday
4. ⏸️ Search delays → Backlog for Q4
5. ⏸️ [Other problems] → Document for later