What do retrospectives look like?

In The Art of Agile Development, Shore and Warden describe a way of holding retrospectives.

They start with Norm Kerth’s Prime Directive (which is enough to make me love Agile, all by itself). Then, they plan out 30 minutes of brainstorming, ending with sticking cards up on a board. After that is ten minutes of a game called “mute mapping”. Folks put the cards together that they think go together, but no one can speak. (I wonder if there are mimed fights over cards, and how funny they’d look.)

Still in that second ten minutes, they progress to grouping the cards into categories, and then voting (by physically approaching the board again) on which categories are the best candidates for improvement.

Once the voting is done, they select the winning category, and put aside all the other cards. (Future retrospectives can deal with those.) During this 20 minutes, some time is put into root cause analysis, and brainstorming ideas for improving things. The group then forms a consensus or votes to determine the one idea they’ll implement during the next iteration.


Being a student, I’ve only been part of one Agile retrospective. For that one, we sat in a circle. We were all pretty happy to be doing what we were doing (teaching & learning Agile-type Java programming in a prison), so we got the happies out of the way. Then we talked about what we might want to shift. People suggested changing things, other people said they liked it the way it was. Vice versa. It was all very laid back. Of course, we were retrospecting a one-day session, in preparation for the next one-day session. Hardly a Serious Agile Project.

So I’m wondering.

What are retrospectives like? Is the Prime Directive really taken seriously? (Oh, please say yes!) Do people do a sort of party-game thing with post-its and magnets and stuff, like in the book? Or are “real” retrospectives more like the one we had in prison? I’m curious!

GeePawHill sez…

Retrospectives take all kinds of form. I sometimes go as simply as writing down goods and bads, identifying the samenesses, and pick one or two that *I* want them to talk about.

Other times, I do things much more as Shore and Warden describe.

Strange as it may seem to say, tho, I have never attended or given a retrospective where we did not mention explicitly that the session was to be bound by the prime directive.

And yes, that does make retrospectives totally cool right from the get-go. I think they are at least a step towards the incredible methodology my friend Steve Doubleday once told me. In Steve’s method, there are only two rules: 1) don’t waste time, and 2) accept the whole person. (Notice that’s not the same as “accept everyone”.)

To me the prime directive reaches toward radical acceptance.

4 thoughts on “What do retrospectives look like?

  1. George Dinwiddie

    Angela, there are many ways of conducting a retrospective. Many Scrum teams perform a perfunctory “What problems do we need to fix” without gaining much insight. Those surface level problems can, and should, be addressed in the day to day work. A good retrospective can take you so much further.

    Still, every team has to start somewhere. Starting with an honest discussion of problems that occurred and what the team would like to do about them isn’t a bad start. It’s perhaps harder to do well than going through all of the stages that Esther Derby and Diana Larsen recommend in their book, Agile Retrospectives.

    I don’t always start explicitly with the Prime Directive, but I certainly do when I think there could be distrust and blaming in the group. I believe strongly in it.

    With a new group, I also often start with a safety exercise, to make more explicit the level of safety that people in the room feel.

    I’ve got some notes and links at http://idiacomputing.com/moin/RetrospectiveTechniques and http://idiacomputing.com/moin/IntrospectionAndRetrospectives that might be of interest.

    The bottom line is that retrospectives are a way to look at the past and use it to choose the future we want. It’s a conscious process, and it’s necessary to vary the details to keep it fresh and conscious.

  2. Darren

    When I’ve done retrospectives with my team, I follow this pattern:

    * Go around the room and get one negative thing from each team member.
    * Go back around the room and let everyone vote. Everyone gets 3 votes to use as they wish.
    * Go around the room (usually the reverse direction) and get one positive thing from each team member.
    * Go back around the room and let everyone vote. Same rules as before.

    Now we pull out the top 3 items from each list and have a short discussion about them. This exercise has helped us focus on the important agile practices and essential elements of teamwork.

  3. Sean McMillan

    I’ve had highly effective retrospectives with just collecting lists of “do more”, “do less”, “do differently”. I’ve had substantially more formal retrospectives that were not helpful at all.

    I think the biggest difference was the facilitator: One was an experienced agile coach who had lots of experience, and a piercing insight into people, which he used to probe for more information. The other was a newly minted scrum master who was having trouble keeping up with a large team.

    Both used the prime directive, but it was not mentioned explicitly. I foud the retrospective much more valuable when it included the creation of actual tasks people could do to change things, and people signing up for those tasks. When no one took ownership of the output, we kept hearing the same complaints over and over.

  4. Angela Post author

    I am appreciating all of these answers. Someday, I’ll have seen a bunch of retrospectives, but for now, stories are what I have to go on. Thanks to all of you for taking the time to answer, and for giving me glimpses into what they can look like.

    @George: Thanks for the links. I enjoyed your focus on creating safety in the beginning. I notice you mentioned that Shore asks for agreement to the Prime Directive in the beginning. I’d add that — and this surprised me — he also recommends cancelling the retrospective if someone can’t agree. Bold! :)

    @Darren: I imagine that I’m not very into structure, but the structure you described seems relaxed and productive. I find that really appealing.

    @Sean: I imagine the attitude and skills of the facilitator make a big difference. When you mentioned the “newly minted” :) Scrum master, something else occurred to me. I bet the facilitator needs to have pretty thoroughly absorbed the Prime Directive internally — that is, also remember not to be blaming herself either before or during the retrospective.


Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>