Sprint_24_retro_notes
Sample write-up. This is what the meeting behind the sample board produced. Replace it with your own after each retro, or delete it if your team prefers to keep everything on the board.
Date: 12 March 2027 ยท Facilitator: Jamie Burke ยท Present: 7 of 7 ยท Format: What went well / didn't / next
The numbers
| Sprint 23 | Sprint 24 | |
|---|---|---|
| Points committed | 31 | 34 |
| Points delivered | 22 | 27 |
| Carried over | 9 | 7 |
| Bugs opened | 14 | 6 |
| Bugs closed | 9 | 11 |
| Stories added mid-sprint | 3 | 2 |
Better than Sprint 23 on every line. Worth noting the commitment also went up, so the delivery gap did not actually close much in percentage terms: 71% then, 79% now.
What the team decided
Three things caused most of the pain, and they are connected. Stories entered mid-sprint. Nothing came out to make room. Work piled up at the end, which is why QA got nine stories in the last two days and why two of them carried over. Treating these as three separate problems would have produced three separate half-fixes.
The team agreed one structural change (the mid-sprint checkpoint) that addresses all three, plus one rule (nothing enters without something leaving) that addresses the cause.
Staging was the other real cost. A day and a half of QA time lost to a full disk. Nobody defended the current setup. Felipe owns the move and the alerting.
Estimates were a symptom, not a cause. The search work was estimated as one lump by people who had not looked at the indexing layer. The fix is splitting the epic, not estimating harder.
What was deliberately not actioned
Standup overrun. Parked. The theory is that the mid-sprint checkpoint absorbs the conversations currently overflowing into standup. If standups are still running long at the Sprint 25 retro, this comes back with an actual action.
WIP limits. Discussed and rejected. The bottleneck this sprint was QA capacity in week two, not too many things started. Revisit if the checkpoint doesn't help.
Rotating engineers into QA duty. Raised in the last four minutes. Deliberately not decided under time pressure. Carried to Sprint 25.
Honest note on the carry-over
Two of three actions from Sprint 23 are still open, and the CI pipeline item has now carried over twice. That was called out in the meeting rather than glossed over.
The CI item is a genuine multi-sprint piece of work that has never been given sprint capacity, which means agreeing it in a retro was never going to be enough. It either goes into a sprint as real work or it comes off the list. That decision is due at Sprint 25 planning.
Actions agreed
Seven, which is more than the three-to-five this template recommends. The team accepted that trade because four of them are under a day's work. If the small ones slip too, the next retro should cut harder.
Full list with owners and dates is in the โ Action items โ Sprint 24 section, and the open ones across all sprints are in the โ Open action items sublist.