Why most retros change nothing
The standard retrospective format — what went well, what didn't, what to improve — was groundbreaking when it was introduced. It is now the default reason retros feel pointless.
The reason: the format produces a list of feelings, not a list of commitments. The team leaves the retro with three "what to improve" bullets and zero owners. By the next retro, the same three issues come up again.
A retro that drives change has two non-negotiable elements:
Every "improve" item ends with a named owner and a deadline
The next retro starts by reviewing whether the previous commitments shipped
If you do not do both of those, you are running a venting session, not a retrospective.
The template
Here is the structure that consistently produces follow-through.
1. Review last retro's commitments (5 minutes)
Open by going through what was decided last time. For each item:
This single change — putting last retro's commitments at the top of this retro — fixes 80% of why retros don't drive change.
2. What worked (10 minutes)
What should we keep doing? Not "what was good" — what was good is feelings. "What should we keep doing" is process.
Examples:
Capture as bullets. Keep moving.
3. What didn't work (10 minutes)
What should we stop doing or change? Same rule — actions, not feelings.
Examples:
Capture as bullets. Don't debate yet.
4. Top 3 changes — with owners and deadlines (15 minutes)
This is the part that matters. From the "didn't work" list, pick the top 3 (not 7, not 12). For each one:
Example:
If you can't pick 3 with owners, you don't have commitments — you have a wishlist.
5. Wrap (5 minutes)
Recap the 3 commitments, by name. Confirm them on Slack so they are written down. Done.
Total: 45 minutes. Anything longer is rumination.
Common retro failure modes (and fixes)
Failure: "We need to communicate better."
Vague. Owner-less. Means nothing. Fix: rewrite as a specific behavior change with one owner. "We will update the project channel every time a blocker is hit, by EOD same day. Owner: whoever hit the blocker."
Failure: One person dominates the retro.
Common. Fix: silent ideation first — every team member writes their bullets in a doc for 5 minutes before discussion. Reduces the "first mover" effect.
Failure: Same issues come up retro after retro.
The single biggest signal that the team is not driving change. Fix: make Step 1 (review last retro's commitments) the first agenda item, religiously. Public accountability drives follow-through faster than any process tweak.
Failure: Retros run long and people get bored.
The team is treating retro as therapy. Fix: enforce the time-boxes above. 45 minutes max. The discipline of a tight box forces the conversation toward decisions.
Capture the retro in 20 seconds, not 20 minutes
The post-retro work is usually a recap email to the team plus an updated commitments doc. Both can be drafted automatically.
While the retro happens, jot down bullets in any format — fragments are fine. Then paste those notes into MeetingFlash. It produces:
That whole post-retro packaging that usually takes the scrum master 20 minutes — done in 20 seconds.
We've also built a sprint retro template into the templates dropdown that prompts you to capture exactly the inputs above.
The compounding effect
Retros that drive change compound. If your team ships even one process improvement per sprint, you are 26 improvements better in a year. Most teams ship zero, because they never close the loop.
The fix is not a fancier retro framework. It is making sure every retro ends with named owners and the next retro starts by checking on them.
For broader team workflow improvements, see The Post-Meeting Workflow That Keeps Teams Aligned.