A retrospective is one of the most valuable rituals a team can have, and one of the easiest to get wrong. Done well, it creates a space where the team can be honest about what is and is not working, and turn that honesty into concrete improvements. Done badly, it becomes a venting session that changes nothing, or a box-ticking exercise that people stop investing in. The structure matters, but so does the intent.
A retro is not about blame and it is not a complaint forum. It is a structured conversation about how to work better, and it only works if it ends with real actions that get real follow-through.
The five phases of a good retro
Regardless of which format you use, effective retrospectives tend to move through five phases. Skipping any of them, particularly the last two, is usually why retros fail to generate change.
- 1
Set the scene
Establish the purpose and create safety. Remind the team that the goal is improvement, not blame. If the team is new to retros, read the Prime Directive aloud: "Regardless of what we discover, we understand and truly believe that everyone did the best job they could..."
- 2
Gather data
Give everyone space to share observations from the period you are reviewing. Use a structured format (see below) so contributions are organised. Silent writing first, then share. This prevents groupthink and ensures quieter voices are heard.
- 3
Generate insights
Group related themes, identify patterns, and discuss what is behind the observations. The goal here is not to problem-solve yet. It is to understand. Ask "why?" more than once.
- 4
Decide on actions
Agree on two to four specific, concrete changes. Each action needs an owner and a timeline. Resist the temptation to tackle everything. A short list of actions that get done beats a long list that disappears.
- 5
Close clearly
Summarise what was agreed. Appreciate the team for their honesty. Check in briefly on how the retro itself felt. A one-word round is enough. Then make sure the actions are recorded somewhere visible.
Choosing the right format
The format you choose shapes the conversation. Some formats are better for teams that are new to retros and need simple prompts; others work well for mature teams that want more nuance. The format matters less than facilitating it well, but choosing one that fits the team's mood and context helps.
- Start / Stop / ContinueThe most common starting format. What should we start doing? What should we stop? What is working that we should keep? Simple, accessible, and good for teams new to retrospectives.
- 4 LsLiked, Learned, Lacked, Longed For. Useful when you want to surface both what went well and what was missing, with a more reflective tone than Start/Stop/Continue.
- Mad / Sad / GladFocuses on emotional response to the period. Useful when team morale or dynamics are a key concern, or when the team needs to process a difficult period before moving to practical improvements.
- SailboatThe team is a boat: what winds are pushing us forward, what anchors are slowing us down? Good for teams that respond well to visual metaphors and want to connect the retro to broader direction.
- TimelinePlot key events from the period chronologically, then discuss what went well and what did not at each point. Particularly effective for longer cycles or projects where the sequence of events matters.
Common mistakes that kill retros
Most retros that fail do so for predictable reasons. Recognising these patterns is the first step to avoiding them.
- No actionsA retro that ends with discussion but no commitments will train your team to see retros as pointless. Every retro must close with specific, owned actions, even if it is just one or two.
- Actions not followed upEven worse than no actions is actions that are written down and then ignored. Open the next retro by reviewing what was agreed last time. This single habit transforms retro credibility.
- Blame cultureWhen feedback focuses on people rather than systems, people get defensive and honesty disappears. Keep observations about processes, decisions, and patterns, not individuals.
- Running too longA retro that runs for two hours with no structure leaves people exhausted and results in lower-quality decisions. Sixty to ninety minutes with a clear agenda is usually sufficient.
- Same voices dominatingWithout deliberate facilitation, retros are easily dominated by the loudest people. Silent writing, round-robin sharing, and dot voting all help surface quieter perspectives.
Turning retro outputs into real change
The retro itself is only valuable if what comes out of it actually changes something. That requires two habits: capturing actions clearly enough to be acted on, and reviewing them at the start of the next retro.
Manager Toolkit's retrospectives feature is built around this: run your retro, gather responses by theme, and connect actions directly to the session so they do not get lost. Tag retro actions with Key Themes so you can see whether certain types of problems keep appearing across sessions. When you run the next retro, previous actions are surfaced automatically so you can see what changed and what did not. Over time, this creates a record of how the team has evolved, and makes it much harder to repeat the same conversations without acknowledging them.
Run your next retro in minutes
Templates, real-time collaboration, voting, timers, and action tracking, all in one place.