Retrospectives done right

Retrospectives are regular team formats in which different perspectives on the collaboration are collected, interpreted and translated into concrete next steps. They follow five phases: Creating space, collecting data, gaining insights, deriving measures and closing. The decisive factor is a framework of trust that enables open sharing in the first place.

Key Takeaways

• Retrospectives without changing topics always produce the same results: If you conduct every retrospective with the same questions at the same altitude, you get the same answers and no new insights.
• Measures usually fail not because of a lack of will, but because they are formulated too abstractly: The concrete next step, such as writing an email or setting an appointment, sets more in motion than a big goal.
• The five phases of a retrospective (Set the Stage, Gather Data, Generate Insights, Decide What to Do, Closing) are not a formality, but prevent the most common mistake: jumping into interpretation too early, before all perspectives are in the room.
• It makes sense to start the design of a retrospective with the third phase: if you first clarify what kind of insights are needed, you can align the check-in and introduction accordingly, instead of the other way around.

What distinguishes a retrospective from a workshop

A retrospective is a format in which a team learns from past experiences of working together and reflects on these experiences together. The core lies in the word “together”: reflection does not take place in one’s own head, but in a space in which several people share their different perceptions.

Sabina Lammert compares this to the image of six blind people and an elephant. One feels a rope, the next thinks a leg is a pillar, a third believes it is a bed. Everyone has a different perspective, and it is only when these perspectives are brought together that a picture emerges that comes closer to the truth.

The difference to a normal workshop lies in two things: the dramaturgy and the repetition. A retrospective follows a well thought-out structure that opens up a space in which people are willing to share honestly. And it recurs regularly. This rhythm builds trust over time.

Why continuous reflection is worthwhile

If you don’t take the time for different perspectives, in the worst case you will make the same mistakes over and over again. Continuous improvement is not an agile ritual in itself, but the prerequisite for becoming and remaining profitable.

The only thing that is constant is the change in the framework conditions. Even a team that works well together over the long term operates in an environment that shifts: politically, technically or due to unexpected events such as a pandemic. Without regular pauses, a team will drift unnoticed in the wrong direction and may not be able to make corrections in time.

A retrospective is therefore something like a team’s health check. And the result may also be: Everything is fine, we’ll carry on as before. Then that is exactly the right conclusion.

Trust is the condition, not the accessory

A retrospective only works in a space of trust. Anyone who shares a perspective must be sure that they will be heard and not devalued for their contribution.

Two agreements underpin this space. The first is the Prime Directive, formulated by Norman Kerth: the basic assumption that everyone has done their best under the given conditions. If something is wrong, the team works on the framework conditions, not on the individual.

The second is the Las Vegas rule. What happens in the retrospective stays in the retrospective. Sensitive information only leaves the room if everyone expressly agrees.

Neither can be enforced. In an initial retrospective, no one shares anything confidential, and that’s fine. Only when the format is cultivated will situations arise over time in which participants show themselves to be vulnerable and say: I made a mistake. If it is then possible to respond constructively, mistakes become visible earlier and the team can respond better.

The five phases of a retrospective

A retrospective follows five phases, as described by Esther Derby and Diana Larsen. This structure is the difference between a meeting with few results and real improvement.

| phase | task | |-------|-------------------|---------| | 1 | Set the Stage | Open the room, let people arrive, create the setting | | 2 | Gather Data | Collect perceptions, let every perspective have its say | | 3 | Generate Insights | Interpret the collected data | | 4 | Decide What to Do | Determine concrete next steps | | 5 | Close the Retrospective | Conclusion, Cooldown, Clarify Confidentiality |

Phase one is about literally shaping the space: Circle of chairs or not, online or on site, flipchart or post-its. The participants should be able to arrive and let go of other concerns.

Phase two is the step that is often skipped. Before anything is interpreted, the team collects the individual perceptions, the rope, the pillar, the bed. Only this collection makes it visible that there is something bigger behind it, the whole elephant.

One of the most common mistakes is jumping into the interpretation too quickly. This is precisely why phase two needs time and space so that everyone can really have their say.

Measures fail if they are thought too big

Measures from retrospectives often fail because they are formulated too big. Those who want to jump directly from the status quo to the target image formulate something so abstract that nobody implements it. Then everyone wonders why nothing happens.

The more effective way is to take the concrete next step. Instead of a gigantic measure, all you need to do is write an e-mail and set an appointment. In very small steps. This attitude comes from solution-focused coaching and noticeably changes the success rate.

A second reason for measures not being implemented lies outside your own sphere of influence. The Circle of Influence helps here: What can the team itself influence and what can it not? If the original measure is outside the team’s sphere of influence, the next step may be to talk to someone who can influence it. This overbandeau game is often unavoidable in organizations.

What to do if the same notes come out every time

If retrospectives always deliver the same results, this is usually due to the same questions always being asked at the same level. If you ask “What went well, where is there room for improvement?” every time throughout the sprint, you get the same surface back every time.

The solution is to focus on different topics. Instead of working through twenty topics per session, the team picks out a single topic, such as meetings or the interface to sales, and goes into depth there. In-depth discussions only arise when the level of detail increases.

If a team has not implemented anything from one retrospective to the next, this can be made a topic of discussion. A retrospective on the measures from the last retrospective clarifies the cause: Was the measure too big? Was it outside the sphere of influence? Was someone missing, was there a lack of knowledge or expertise?

It is also important to check in each subsequent retrospective whether the measures taken have been implemented. If you only come up with new measures without checking the effectiveness of the old ones, you are going round in circles.

From format to design: start with the goal

The selection of the appropriate method does not start with the format, but with the goal of the retrospective. Only then does the question follow as to which exercise supports this goal. Collections of methods such as Corinna Baldauf’s Retromat provide numerous formats for each phase, but not every format is suitable for every goal.

Sabina Lammert has developed a design canvas for this, intended as an entry-level aid, not for professionals with hundreds of moderated retrospectives. The approach reverses the usual sequence: The design begins with phase three, with the insights that are to be achieved. The rest is built backwards from this target image.

The model comes from yoga. There is a peak pose that requires particular strength or flexibility. If you know what you are working towards, you build up the warm-up accordingly so that nobody gets injured. Applied to retrospectives, this means that the start is based on the desired result.

The canvas distinguishes between four characters of retrospectives, each of which requires a different approach:

• Analytical: This is about understanding data. The check-in addresses the analytical part of the brain.
• Emotional: There is a conflict or different perceptions of a situation. The check-in is low-threshold and helps you to connect with your own feelings.
• Creative: Solutions should emerge outside the usual channels. The first phase then also starts creatively.
• Hands-on: The team needs to move quickly into implementation and work associatively rather than cerebrally.

If I know what I’m working towards, I have to adjust my warm-up accordingly. At some point, this has become intertwined with the way I design my retrospectives.

Sabina Lammert

Always adapt the method to the team

It’s not just the goal that counts, but also who you’re dealing with. Some teams don’t want to jump around in circles or work with Lego, and this caveat should be included in the design.

Lego is a good example of a method that has been burned in many organizations because of poorly done Lego workshops. Even a convinced Lego fan should then resort to a less playful, more analytical method so that the participants do not start with reservations.

As an external moderator, Sabina gets a feel for this through preliminary discussions. Helpful questions: Which workshops have already taken place? What went down well, what didn’t? Those who know the team themselves have an advantage here and know what has worked in the past.

The way forward is to feel your way around. From retrospective to retrospective, more can be ventured, a little game, a visual exercise, a motto. On the other hand, starting with a completely over-excited motto retrospective with an unknown team with negative previous experiences can easily go wrong.

How to get started as a beginner

Anyone starting out with retrospectives should consciously reduce the complexity. The number of possibilities from five phases, countless formats and free combinations is overwhelming at the beginning.

Three things will help you get started:

• Moderate with someone experienced. Taking a look at the procedure significantly lowers the barrier to entry.
• **Go into sparring: There is often someone in your own organization with experience with whom it is worth exchanging ideas.
• **Participate yourself **Having experienced a really good retrospective once makes a huge difference.