Key takeaways
- An empty retrospective almost always comes from a vague framework, not from a lack of ideas.
- Five formats cover most situations: starfish, 4Ls, mad/sad/glad, Start/Stop/Continue, and the "one action" format.
- Each format dictates its questions; asking them in order prevents debates that go in circles.
- The real pitfall isn't the discussion, it's the disappearance of actions after the meeting.
- Recording each action as a card with an owner and a due date makes it verifiable from one sprint to the next.
Why your retrospective falls flat
A retrospective that runs out of steam rarely has a participant problem. It has a framework problem. Without structure, the discussion drifts toward the complaints of the moment, endless technical topics, or polite silence.
The format isn't a decorative ritual. It's what forces everyone to produce a precise answer rather than a general opinion. A good format fits one question per box, and gets filled in ten minutes.
Second common cause: the previous retrospective changed nothing. If the actions decided last month are nowhere to be found, nobody believes in today's. Tracking matters more than the format.
5 retrospective formats, with the questions to ask
1. The starfish
Five branches, five questions. It's the most complete format when the team needs to frame both what works and what's blocking.
- Continue: what do we keep as is?
- Stop: what do we drop?
- Start: what do we try next sprint?
- More of: what do we want to amplify?
- Less of: what do we want to reduce?
Pitfall: the "more of" and "less of" branches often get filled with the same ideas as "continue" and "stop." Ask for a concrete example for each sticky note, otherwise the box is useless.
2. The 4Ls
Four columns starting with L, useful when the team mixes feelings and facts.
- Liked: what did you like?
- Learned: what did you learn?
- Lacked: what was missing?
- Longed for: what did you wish for?
Pitfall: "Longed for" quickly becomes a list of material wishes. Refocus on the work: what tool, what information, what decision would have saved you time?
3. Mad / Sad / Glad
Three columns, three emotions. A quick format, suited to teams of 2 to 6 people or short sprints.
- Mad: what frustrated you?
- Sad: what disappointed you?
- Glad: what satisfied you?
Pitfall: emotion takes up all the space and no action comes out. Block the last five minutes to turn each "mad" into a verifiable proposal.
4. Start / Stop / Continue
The most direct for a team that already knows its problems and wants to decide.
- Start: what should we start doing?
- Stop: what should we stop doing?
- Continue: what should we keep doing?
Pitfall: "Stop" becomes a trial. Remind the rule: we talk about practices, not people.
5. The "one action" format
When the team is saturated with meetings, reduce to a single question: what is the one thing that, if we change it this sprint, most improves the way we work? One answer, one action, one owner. Ten minutes.
Pitfall: choosing an action that's too broad ("communicate better"). Demand an observable criterion: who does what, and how we'll know it's done.
The pitfalls that come back in every format
Three mistakes come back regardless of the format chosen.
- Too many sticky notes. Beyond fifteen contributions, the group no longer prioritizes. Group them, then vote.
- No owner named. An action without a name is a wish. Write the first name on the card, not "the team."
- No due date. "Next sprint" is enough, as long as it's written down.
One last point: don't turn the retrospective into a sprint review. Bugs and delays are handled elsewhere. Here, we talk about the way we work.
Recording actions so they don't get lost
The retrospective produces two things: observations and actions. Observations can stay in a report. Actions cannot: they must live in the same place as the work.
The simple method: create one card per action, with a title in the infinitive ("Reduce pull request size"), an owner, a due date, and a verification checklist. Place these cards in a dedicated column of the Kanban board, or attach them to the current sprint.
Ever Earlier allows exactly that: each card carries an owner, a checklist and an activity log. At the next retrospective, you open the card and see who did what, when, and what remains. No reconstructing from memory, no debate about what had been decided.
Two rules to make it hold:
- Limit yourself to one to three actions per retrospective. Beyond that, nothing gets done.
- Reread the action cards at the start of the next retrospective, before adding new ones.
The activity log also serves as memory for the team: a comment at the moment the action progresses is worth more than a report rewritten three weeks later.
Checklist before launching your next retrospective
- A format chosen, displayed, with its questions written out.
- A slot of 45 minutes maximum, timed.
- A rule reminded: we talk about practices, not people.
- Time dedicated to turning observations into actions.
- One to three actions, each with an owner and a due date.
- Each action recorded in a card, not in a separate document.
- The previous sprint's actions reread at the opening.
Sources
Read next
WIP Limits: The Number That Changes Your Flow
A poorly calibrated WIP limit turns your Kanban board into a silent queue. Here is how to calculate it, test it over two weeks, and measure its effect on lead time. Per-column calibration exercise included.
Ineffective Daily Standup: The 3-Question Test
Your daily standup drags on, goes in circles, and no one dares say it. In one week, three questions are enough to decide: coordination or theater. Here is the diagnostic protocol, the thresholds to measure, and the asynchronous format for distributed teams.
Kanban or Scrum: Choosing the Right Method in 2026
Continuous flow or iterations? The choice depends mainly on your delivery frequency and the nature of your product. Here are concrete criteria to decide, and how to configure each approach in a tool like Ever Earlier.