Following the discussion I had in my head with Aino Vonge Corry during her talk at LeanAgileScotlnaid in 2022, I wanted to throw on the paper the union of our best practices. (And have a look at her book on Retrospective Antipatterns!)

Setting the stage should also be the moment where everyone has to speak, in order to “activate the right to speak”.

I do it by forcing people to say one top and one flop, and giving an exemple that they can copy, just so they can hear the sound of their voice.

Gathering data can be something else than just writing on post-it. But gather problems, not solutions. Neither symptoms. Put them on the side.

That’s why I don’t like “keep drop start”, but use “top, flops, action” saying that we’ll talk about actions after everybody will have given their tops/flops. If an action pops indeed I might keep it if it’s important for the participant.

Have a root cause analysis somehow, to generate insights. Through a fish bone diagram for instance. And remember that there is full spectrum of preference between global thinkers and sequential thinkers: you might not want to put labels to all your bones.

I try to go into causal analysis after the data is gathered. Moreover, it’s really hard to see the difference between digging to find the objective fact, and inventing hypothetical cause-consequence relationship. I thus ask for examples in this case.

The groan zone is a really interesting zone: needs psychological safety to disagree; and sometimes it’s a zone the team have the culture to get out from this zone. But it might be another zone that your team hate.

Assume that you can’t change the soup! Yet, there are things you can bring closer from the soup.

I use the French expression “balayer devant sa porte”. Don’t formulate actions starting with “they could/should” but “we could/should”.

About prime directive: can be controversial, we ourselves know that we have not done our best. The point here is to forgive.

Manage loud mouths, but with respect, allowing them to save the face. They might not know they are. It might be a valorised posture in their communities.

A facilitator should not be doing things.

That’s what I always begin with my points so that I switch to the facilitator mode asap.

For online retrospectives, prepare it! To compensate lack of body language, capabilities to change easily the support…

And that’s it!

Not for me :) ! I missed something about self-censorship, and trust and censorship in general. And the right to fumble their words for: some people need to speak to think, and thus that the first oral draft will be imprecise or even false. Precising that I always give 3-5 minutes at the beginning of pure silence for those who need to think before speaking to write their ideas on a paper.

  • Note 1: antipatterns solve the problem in a way that is has to be refactored the after
  • Note 2: you can sell retrospective to management saying the team will learn. But its a place to share and appreciate too.