The 5 Phases of the Retrospective

Why not to talk about important stuff and then just decide what to do?

Why not ask the team if they find the phases you describe useful? It's their Sprint Retrospective, how to implement it is a matter for them, they own it.

thx but that was not my question, my question was if somebody does it this way, what is the purpose of each phase? And how to connect the phases, so that it makes sense?

It sounds like you're referring to the five stages of a retrospective that are described in Esther Derby and Diana Larsen's Agile Retrospectives: Making Good Teams Great. The five stages are setting the stage, gathering data, generating insights, deciding what to do, and closing the retrospective.

What isn't clear to me is what you mean by "the transition between phases". I don't think the phases are discrete blocks of time. That is, I don't think that the retrospective facilitator is telling the team that it's time to set the stage and then time to gather data and so on. Instead, the stages are a natural set of things that need to happen in order for a team to achieve the desired outcomes of a retrospective. I'd suspect that most good retrospectives follow these stages, even if the team or facilitator is unaware of them I don't think that you can just "talk about the important stuff and then decide what to do" without going through the phases.

Let's look at an example, though.

Setting the stage is straightforward and gets everyone in the mindset to perform a retrospective. For some teams, this can be a good opportunity to remind everyone that the retrospective is a safe space. The things that people say or talk about shouldn't be brought up outside of the group or attributed to individuals.

Gathering data is where you talk about the important stuff. Some data could be related to the team's velocity and throughput for the Sprint, but it doesn't have to be quantitative data. Understanding events that happened in the Sprint and how the team felt about them is just as important. There's plenty of qualitative data out there for the team.

Not all of the data is actionable, though. Of all the data available, the generating insights portion is about making sense of it. If the team's throughput is consistent with recent Sprints, maybe that's not something the team wants to look at. Perhaps the team is feeling more stressed recently, that could be something to dive into. It doesn't have to be negative, either. In cases where the team does really well, they may want to dive into what they can do to repeat that. Generating insights is about getting to the root causes and contributing factors to one or two of the most important things that happened to the team.

Once you identified the important stuff, then you can decide what to do about it. Taking time to get to the root causes can help you build better plans for actually solving problems or making successes more repeatable. It's not a random guess at what to change, but very specific, targeted changes to the team's way of working to increase success in the future.

Just about every retrospective I've participated in follows this type of approach. They may call the stages different things, such as brainstorming, root cause analysis, and action item planning. Some of them may be asynchronous - some teams I've worked with use a Confluence page or Google Form to let team members capture events and thoughts (or gather data) throughout the Sprint and focus on generating insights and developing a plan of action at their retrospective. Even so, all of the activities happen.

@Thomas Owens just gave the best summary that I have ever read.  Since he did such a good job, I'll provide an example of how I have facilitated a large number of retrospectives for teams that are new to the concept. This is only a single example and there are MANY ways to do this.  I will often change the format so that people don't get complacent and variety can help people engage. 

---------------------------

Set the stage

I start by explaining that the purpose of the gathering is to openly discuss anything that is troubling or helpful to each individual.  There are no right or wrong answers and everyone is free to provide their own opinion.  The intended goal is for the team to agree on a few things and provide ideas to either avoid or continue those things. I remind everyone that anything we agree to do will be considered an experiment and can be accessed/discussed at the next retrospective. I explain the concept of "disagree but support" because not everyone will get exactly what they want.

I have used this to set the stage on occasion. https://retrospectivewiki.org/index.php?title=The_Prime_Directive

Gather Data

I provide some sort of mechanism for the participants to "rush the board". In the old days, that would be everyone standing up, walking to a whiteboard/flip chart with a pad of sticky notes and a sharpie, and taking # minutes to write down anything they want with each item on a separate sticky note, and put it on the board.  This gives everyone a chance to think on their own and provide their thoughts without having to wait their turn.

Generate insights/Decide what to do

After everyone has put their items on the board, I will facilitate a discussion to group similar items together.  After the grouping I ask everyone to "dot vote".  Dot voting is when everyone returns to the board, using their sharpie they place a dot on sticky notes.  Each individual gets X votes, I typically use 3, that they can place in any way that they want. They can place them on 3 items or all 3 on a single. It is their choice how they want to use them.

After voting is complete, I will facilitate discussions starting with the groups that have received the most votes. The discussion is focused on coming to an agreement on what the team would like to do about the items that were grouped together. When a decision is agreed upon, it will be written on a sticky note and placed on top of that group.  I then move to the next group.  

We also discuss how we will surface the decisions and how the team will hold each other accountable for abiding by the decisions.

I honor the time limit and try to cover as many groups as possible.

Closing the retrospective

I thank everyone for their contributions, remind them of how they have decided to surface and hold each other accountable.  And then I remind them that the retrospective is not the only time that these discussions can happen.  

-------------------------

Variety and practice are beneficial. I often use Corinna Baldauf's Retromat which randomizes a selection of techniques for each activity.