translate Automatic translation · View original
Retrospectives that matter: 10 formats that bring continuous improvement back to life
In their book Agile Retrospectives: Making Good Teams Great, Esther Derby and Diana Larsen define the retrospective as "a dedicated opportunity for the team to inspect and adapt its processes and practices". However, in practice, many retrospectives have been reduced to a mechanical formula that generates more frustration than improvement.
The four symptoms of an exhausted retrospective
1. Lack of psychological safety
This is considered one of the most important factors for team effectiveness (Edmondson, 2018). When members do not feel safe expressing real problems (especially those related to power dynamics, decisions, or interpersonal conflicts), the retrospective becomes a theater where everyone acts as if everything is fine.
2. Absence of real follow-up
It often happens that retrospective commitments are rarely fully implemented. This creates a vicious circle: the team proposes improvements, nothing changes, trust erodes, participation decreases, and the next retrospective is even less productive. When retrospectives have no real consequences, the implicit message is devastating: "your opinion doesn't really matter".
3. Repetitive and predictable format
When a team does the same thing every time, it stops seeing what it has not seen before. Repetitive formats generate automatic responses. The brain goes on autopilot and genuine reflection disappears. If you have been using the same format for six months, your team already knows exactly what is going to happen. And when we know exactly what is going to happen, we stop paying real attention.
4. Disconnect between identified problems and capacity to act
Perhaps the most frustrating problem: the team identifies systemic impediments it cannot solve. They need another area to respond faster, a budget approved for tools, or an organizational policy revised. But the scrum master has no authority to change that, and the team feels powerless.
Henrik Kniberg points out that "one of the biggest demotivators is identifying problems you know you can't solve" (Kniberg, 2014). Effective retrospectives require distinguishing between what the team can change directly and what needs to be escalated, and having clear paths for both cases.
What makes a retrospective truly matter
Before jumping to formats, we need to understand what makes a retrospective more than an event on the calendar.
Clear and contextual purpose
Not all retrospectives should address the same topics. For Patrick Lencioni, teams evolve through different challenges in the following order: trust, productive conflict, commitment, accountability, and results. An effective retrospective diagnoses what stage the team is in and designs the conversation accordingly.
Is your team dealing with an unresolved conflict? A retrospective about "what went well and what to improve" will not help. You need a format that allows difficult conversations. Have they just launched an important feature? The moment calls for celebration and learning, not process optimization.
Intentional variation
Research on creativity and problem solving consistently shows that changing context generates new perspectives (Guilford, 1967). Changing the retrospective format is not entertainment: it is a cognitive tool. Different formats activate different types of thinking and allow us to see problems from angles the usual format hides.
From insight to action
For Lyssa Adkins, "reflection without action is philosophy; action without reflection is recklessness". An effective retrospective does not end with a list of problems or a board full of sticky notes. It ends with one or two specific commitments, with clear owners, and with an explicit follow-up plan at the next retrospective.
Safety as a prerequisite
Without psychological safety, there is no real retrospective. This means the scrum master needs to actively work to create and maintain a space where it is safe to tell the truth. Sometimes this requires prior one-on-one conversations. Sometimes it means using formats that allow initial anonymity. Sometimes it involves having difficult conversations about behaviors that erode safety.
10 formats that bring continuous improvement back to life
These formats are organized according to the team's context and need. They are not magic formulas: they are tools that work when used with intention and adapted to your specific context.
1. Sailboat
When to use it: when the team needs a visual and metaphorical perspective that allows less confrontational conversations about impediments.
How it works: Draw a sailboat on the sea. The boat represents the team. Identify visual elements:
- Wind in the sails: what pushes us forward.
- Anchors: what holds us back or slows us down.
- Rocks/reefs: future risks we see on the horizon.
- Island/destination: where we are going (the Sprint Goal or the project goal).
Participants add sticky notes to each element. The metaphor allows talking about difficult problems ("this anchor is holding us back") in a less personal way than "Juan always arrives late to the Daily Scrums".
Source: this format was popularized by Innovation Games (Hohmann, 2006).
2. Timeline retrospective
When to use it: at the end of a project, a major release, or after a particularly intense sprint where many things happened.
How it works: Create a visual timeline of the sprint or project on the wall (physical or virtual). Divide it into three rows:
- Events: what happened (deploys, key meetings, decisions).
- Emotions: how the team felt (it can use emojis or colors).
- Learnings: what we understood at each moment.
The team walks through the line chronologically, reconstructing the shared story. This makes it possible to identify patterns ("we always get stressed at the end of the sprint") and correlations ("our mood dropped after that meeting with stakeholders").
Source: Derby, E., & Larsen, D. (2006). Agile Retrospectives: Making Good Teams Great. Pragmatic Bookshelf.
3. The conversation café
When to use it: when you need to go deeper into a specific topic the team has identified as critical, but which requires more time and reflection than a standard format.
How it works: Divide the team into small groups (3-4 people). Each group discusses a specific question for 15-20 minutes. Afterwards, people rotate to other groups, carrying insights from the previous conversation. The process is repeated 2-3 times.
Example questions:
- "What makes this team a place we want to work in?"
- "If we could change just one thing about how we work, what would have the greatest impact?"
- "What conversation do we need to have but have been avoiding?"
In the end, each group shares a synthesis of the most important insights.
Source: Brown, J., & Isaacs, D. (2005). The World Café: Shaping Our Futures Through Conversations That Matter. Berrett-Koehler Publishers.
4. Lean Coffee Retrospective
When to use it: when you want to give the team full control over the topics to discuss, especially if there are multiple pending topics or diverse interests.
How it works: Adopt the Lean Coffee format for the retrospective:
- Topic generation (5 min): each person writes topics they want to discuss.
- Dot voting (3 min): everyone votes for the most important topics.
- Timeboxed discussion: you start with the most voted topic. Set a timer (5-8 minutes). When time runs out, the team does "thumbs up/down/sideways" to decide whether to continue with that topic or move on to the next one.
- Action capture: someone documents specific actions that come up.
This format ensures that time is invested in what the team considers most valuable at that moment.
Source: Lean Coffee is a format created by Jim Benson and Jeremy Lightsmith.
5. Starfish
When to use it: when the Start-Stop-Continue format feels limiting and you need more nuance in the answers.
How it works: Draw a five-pointed star. Each point represents a category:
- Keep doing: what works and we should maintain.
- Less of: what we should reduce (not eliminate completely).
- More of: what we should do more frequently or intensely.
- Stop doing: what we should stop doing.
- Start doing: what new practices we should try.
The difference from Start-Stop-Continue is granularity. "Fewer meetings" is different from "eliminate meetings". "More pair programming" acknowledges that we already do it, but we want more.
Source: this format was popularized by Patrick Kua.
6. 4Ls: Liked, Learned, Lacked, Longed For
When to use it: when you want to combine retrospective and learning review, especially useful after learning a new technology or practice.
How it works: Four categories:
- Liked (I liked): what worked well, what you enjoyed.
- Learned (I learned): what new knowledge or skill you acquired.
- Lacked (It was missing): what we lacked or what obstacles we encountered.
- Longed for (I wished for): what you would have wanted to have or to happen.
This format balances celebration (liked), learning (learned), and improvement (lacked/longed for). It is particularly good for teams that tend to focus only on problems.
Source: Derby, E., & Larsen, D. (2006). Agile Retrospectives: Making Good Teams Great. Pragmatic Bookshelf.
7. Speed Dating Retrospective
When to use it: for large teams (8+ people) where some people dominate the conversation or where there are subgroups that do not interact regularly.
How it works: Organize the team into pairs. Each pair has 5 minutes to discuss a specific question. Afterwards, people rotate to form new pairs with a new question. Repeat for 3-4 rounds.
Example questions:
- "What is the most frustrating thing about how we currently work?"
- "What experiment would you like us to try next sprint?"
- "What worries you that nobody else seems to notice?"
In the end, each person shares the most important insight they heard (not their own opinion, but what they learned from others). This encourages active listening.
Source: adapted from facilitation techniques of Liberating Structures (Lipmanowicz & McCandless, 2013).
8. KALM (Keep, Add, Less, More)
When to use it: similar to Starfish, but with slightly different language that some teams find more intuitive.
How it works: Four quadrants:
- Keep: practices that work and we want to preserve.
- Add: new practices we should experiment with.
- Less: things we do too much of or should reduce.
- More: things we do but should intensify.
The advantage of KALM over other formats is its balance: it explicitly recognizes that not everything is binary (do/don't do) and that many practices need calibration rather than elimination.
9. Constellation
When to use it: when you need to understand the distribution of opinions in the team on specific topics, especially useful for identifying hidden disagreement.
How it works: Designate a physical space (or a virtual one with positioning on Miro/Mural). The facilitator makes a statement: "I feel safe expressing dissenting opinions on this team". Participants position themselves in the space according to their level of agreement:
- Near the center = Totally agree
- Far from the center = Totally disagree
The spatial visualization instantly reveals whether there is consensus, polarization, or distribution. Afterwards, some people from different positions are invited to explain their perspective.
Example statements:
- "We have clarity about our priorities".
- "Our ceremonies are productive".
- "We feel we can influence how the team works".
Source: based on systemic facilitation techniques and organizational constellations.
10. Pre-Mortem Retrospective
When to use it: at the start of an important project or critical sprint, not at the end. It is a proactive retrospective.
How it works: Gary Klein (2007) developed the "pre-mortem" technique to improve decision making. Adapted to retrospectives:
- The facilitator poses: "Imagine it is at the end of a sprint/project and it has been a total failure. What went wrong?"
- Participants, assuming the failure has already occurred, individually write all the reasons that caused it. Afterwards they are shared and grouped by themes.
This technique takes advantage of "hindsight bias" in our favor: when we assume something has already happened, we are much better at identifying potential causes than when we try to predict the future.
The team then prioritizes the identified risks and creates mitigation strategies.
Source: Klein, G. (2007). "Performing a Project Premortem". Harvard Business Review.
What to do next
Changing the format is not enough. These ten formats are tools, but their effectiveness depends on how you use them and what you do with the results.
Bob Galen (2013) proposes a "Definition of Done for Retrospectives": a retrospective is not "finished" until:
- One to two specific actions have been identified (no more; depth is better than breadth).
- Each action has a clear owner (not "the team").
- It has been defined how we will know whether it worked (metrics, observations).
- Time has been reserved in the next sprint to implement the actions.
- The next retrospective starts by reviewing what happened with the previous actions.
Without this follow-up cycle, even the most brilliant retrospective becomes a catharsis session with no impact.
Escalating impediments the team cannot solve
Not all problems identified in retrospectives are within the team's control. The scrum master has the responsibility to take these impediments to whoever can resolve them (Schwaber & Sutherland, 2020).
This requires:
- Clear documentation of the impediment, its impact on the team, and why it cannot be resolved locally.
- Established escalation paths (who do I take this to?).
- Transparency with the team about progress ("I took this to X, it's in progress" vs. "it was blocked for Y reason").
- Honesty about limits (sometimes the answer is: "this won't change in the short term, how do we adapt?").
Cultivating psychological safety continuously
Psychological safety is not built in retrospectives; it is built every day in how the team responds when someone makes a mistake, questions a decision, or admits not knowing something.
Edmondson (2018) identifies specific leader behaviors that foster psychological safety:
- Frame the work as learning problem, not execution problem: "We are discovering how to do this" vs. "We just need to execute the plan".
- Acknowledge your own fallibility: as scrum master or Product Owner, admitting when you don't have answers.
- Model curiosity by asking lots of questions: instead of giving answers, asking "what do you think?" or "what else should we consider?".
Retrospectives are a thermometer of the team's psychological safety. If nobody talks about real problems, the problem is not in the retrospective format: it is in the team's culture.
When the format is not the problem
Sometimes, exhaustion with retrospectives is not solved by changing the format. It may be a symptom of deeper problems:
The team is not really a team
If people work on completely independent tasks with little interdependence, the retrospective feels artificial. Continuous improvement requires joint work to optimize. This is a team design problem, not a retrospective format problem.
The scrum master facilitates but does not lead change
Facilitating the conversation is important, but if the scrum master is not actively removing impediments, following up on actions, and challenging the status quo, retrospectives become venting sessions without consequences.
The organization does not really value continuous improvement
If the organization's implicit message is "optimize your process, but do not question decisions made above", the retrospective has an artificial ceiling. Teams quickly perceive this hypocrisy and disengage. This requires difficult conversations with management about what level of autonomy and authority the team really has.
What makes your retrospectives matter or stop mattering? We would love to hear your experience.
Bibliography
- Brown, J., & Isaacs, D. (2005). The World Café: Shaping Our Futures Through Conversations That Matter. Berrett-Koehler Publishers.
- Derby, E., & Larsen, D. (2006). Agile Retrospectives: Making Good Teams Great. Pragmatic Bookshelf.
- Edmondson, A. C. (2018). The Fearless Organization: Creating Psychological Safety in the Workplace for Learning, Innovation, and Growth. Wiley.
- Galen, B. (2013). Scrum Product Ownership: Balancing Value from the Inside Out (2nd ed.). Outskirts Press.
- Guilford, J. P. (1967). The Nature of Human Intelligence. McGraw-Hill.
- Hohmann, L. (2006). Innovation Games: Creating Breakthrough Products Through Collaborative Play. Addison-Wesley Professional.
- Kerth, N. L. (2001). Project Retrospectives: A Handbook for Team Reviews. Dorset House Publishing.
- Klein, G. (2007). "Performing a Project Premortem". Harvard Business Review, 85(9), 18-19.
- Lencioni, P. (2002). The Five Dysfunctions of a Team: A Leadership Fable. Jossey-Bass.
- Lipmanowicz, H., & McCandless, K. (2013). The Surprising Power of Liberating Structures: Simple Rules to Unleash A Culture of Innovation. Liberating Structures Press.
- Schwaber, K., & Sutherland, J. (2020). The Scrum Guide: The Definitive Guide to Scrum: The Rules of the Game. https://scrumguides.org
- Scrum.org (2023). State of Scrum Report. https://www.scrum.org/resources/state-scrum-2022-2023
Take part in the poll
This article includes an active poll in the Scrum Manager community.
poll View the poll in the community