Every team's first retro is good. People are honest, a little relieved to finally say the thing out loud, and a few real changes come out of it. The fifth one is fine. By the fifteenth, most teams are just filling in the same three sticky-note columns out of habit, and the meeting quietly becomes the thing people tolerate rather than use.
The problem usually isn't the team. It's the format. "What went well / what didn't / what will we do differently" is a good default, but running the exact same three questions every two weeks for a year guarantees diminishing returns — you eventually run out of new things to say inside that specific frame, even though plenty is still worth discussing.
Rotate the lens, not just the questions
A simpler fix than inventing new questions each time is rotating between a small set of different lenses, so the team's attention actually moves:
- The process lens — start/stop/continue, focused only on how work moved through the sprint.
- The energy lens — ask what gave people energy this sprint and what drained it. This surfaces burnout and morale issues the process questions never touch.
- The one-thing lens — skip the categories entirely. Everyone writes one sticky note: the single thing they'd change if they could only change one. Vote, then spend the whole retro on that one item instead of skimming six.
Three lenses, rotated every retro, means no format gets stale before it comes back around.
Track whether anything actually changes
The fastest way to kill a retro's credibility is running one, generating action items, and then never mentioning them again. People notice. After a few cycles of that, the honest feedback dries up because it clearly isn't going anywhere.
Keep the physical tools boring and reliable
This sounds minor, but it isn't: half-dead markers and sticky notes that peel off the whiteboard mid-meeting eat real time and break the flow. A retro where people are re-sticking notes to the wall every two minutes is a retro where attention is somewhere else. Cheap, working supplies aren't glamorous, but they're one of the few retro problems that's actually easy to solve outright.
When to stop running them the usual way
If a team has been stable for a long time and retros are consistently short and low-signal, that's not automatically a problem to fix — it can mean the team has genuinely worked through its recurring issues. In that case, dropping to a monthly cadence, or switching to an as-needed retro triggered by a rough sprint, is often more honest than forcing a biweekly ritual that has nothing left to surface.