I want to share with you my first experience using a retrospective dialogue sheet (the sheet was made by Allan Kelly, I just translated it in French).
Context : I begin 3 weeks ago coaching a web developement team in a French well known mutual health insurance company. The team (and the company) are completely new to Agile.
The manager of the service decided some month ago to “go agile”. It started with another team, who embarked an Agile coach to guide them. The team I am just starting to coach has already begun to borrow some Agile practices, but I would say that for the moment, besides the use of a Scrum board and sticky notes, they are not yet agile.
But, and that is the most important, the spirit is already there : solid team, skilled team members, eager to learn more about Agile, and already producing high quality software.
I had attended their last sprint retrospectiwe, which was quite nice, but I thought I had to show them something else.
As they had just finished the first release of a project, I decided we could make a release retrospective.
So I translated the T1 retrospective dialogue sheet, printed it in A1 format for 8 euros at one printer in the town where I live, took the sheet with me in the train, and met with the team the next day (I coach them 2 days per week).
The crew was :
The 3 developers, the scrum master, me, and the would-be new product owner (the current one was not really the adequate person, we thought it was time for a change).
The whole team sat around a round table, I put the sheet on the table, explained very briefly the purpose and the rules, and there we go !
I was sitting very close to them. I decide to stay during this retrospective for several reasons :
- I had to learn more about their context
- I was curious to see how it would work , and what I could improve to facilitate such a retrospective
- I thought I could guide and facilitate a little this retrospective, because they have to learn agile, and solve important impediments (find an empowered and available product owner being the most important).
What was fine :
- They learned very rapidly how to use the sheet, how to read the questions in turns.
- Everybody felt concerned.
- They took it very seriously: it looks like a game, but people are not playing, they really work.
- Some who usually do not speak much were forced to speak because each one has to speak in turns.
- It was interesting to see them fill the important events on the timeline, trying to remember precisely the launch date, the date of the first stress test, the first demo, the date where the first worries about performance appeared, the time when they solved those problems.
Interestingly, they turned to me at one moment to ask if the start date should be the date when the proper programming started, or the date when they first started working on the POC. That was interesting to notice that they wondered if they were doing it right, and asked me what to do. I told them : do as you feel. So they decided collectively. Interesting how people always think about rules, what to do and what not to do….
They decided that the retrospective should last 1:15 hour. It lasted much more (2:30) but there is a very good reason for that. When we got to the point of “list the main things to do differently”, I jumped to the board, and started a root-cause analysis diagram, to find the cause of the main impediments that they had listed during this retrospective :
- difficulty to track wat was still to be done druing the release
- no real product owner
- too much testing at the end
By doing the root cause analysis, we quickly decided what to do differenlty, what to improve, but we were focusing on the root cause and not just the symptoms.
- People liked to use different colours to write on the board.
- The discussion was calm, everybody spoke, everybody felt concerned, and they spontaneously decided when to go to the next question on the sheet.
What to improve :
- I should have displayed the countdown timer on a big computer screen visible by everyone. In fact one person was using her iPhone countdown timer, so he was the only one to see it, and did not check very often.
- The team agreed that in addition to timeboxing the whole sheet, it would have been good to timebox each question. That is the case in other timesheets, we just have to do the same on this one.
What to try
Perhaps next time I will leave the room and leave the team completely alone during the retrospective, so that there is not that censor type of person. Perhaps the discussion will be more “intimate”.
At the end of this retrospective the whole team agreed that they should use this dialogue sheet again!