Friday, November 29, 2013

My Retrospective About My Retrospective

Business as (too) usual

Back when I was working for my previous employer, we used SCRUM as our development methodology. As part of this methodology we used to end every sprint (3 weeks of work) with a retrospective.

During the retrospective, the team would sit down together and discuss what had been done well and what had not.

This was a recurring ceremony in which the team-leader guided us into talking about the painful pain points (along with the good things we've done - we don't want to be ALL negative). We used to discuss various issues; a lot of those times it was done with a lot of enthusiasm, maybe even some arguing.

Let's make things more exciting

During one of our 1x1s, my team-leader told me that he felt like we had to shake the dust from the retrospective process a bit, make it more exciting and, more importantly, more productive. His suggestion was to have a rotation between team members and let each of us lead a retrospective. We were encouraged to freshen things up, any way we deemed fit.

What should I do?

I started thinking about what I wanted to do when it was my turn to lead a retrospective. Should it be around one specific topic? Should it be a game? Should I stick with the original format?

I decided to think about what was bothering me in our retrospective process. It was a sort of a retrospective with myself about our retrospective.

I narrowed it down to 3 issues:
  1. Priorities: I realized that we don't always address the most important issues. We let our emotions, instead of our common-sense, affect the topics we discuss. Instead of finding 1 or 2 things that can improve us best, we discussed things that really annoyed us. We could not always change these things , or know that the change would have a great impact on the team or productivity, even if it happened.
  2. Not everyone's voice can be heard: Some people are more passionate when they speak, some are shy and some are more verbal and can convince others. It's hard to let everyone be equally heard and it was really important for me to try.
  3. Action items: We all want to change the world, or at least our organization, for the best. But we have to be realistic and set feasible goals that, in each sprint, would improve us a little more. 
So I decided to create a flow in my retrospective which will allow everyone to speak their minds, and will lead to an extremely focused set of 2-3 feasible and improving action items.

So that's what I did

Step 0: Preparations

I booked a room with a big erasable board. Before the team arrived I wrote down all of the stories we worked on during this sprint with different colors (in order to make it very clear where we were at the end of this sprint):

I put some papers and pens on the table for the team to use later on.

This is how the board looked like when the team arrived

Step 1: Map the problems

I asked the team (myself included) to take 3-5 minutes and write down the four impediments which bothered them most. After all were finished we took turns in reading our notes and then wrote each of those on the board, near the stories that were already written there. Everyone could speak their minds about what was bothering them, without interruptions.

In Step 1 we added our painful impediments

Step 2: Suggest solutions

Pointing what's wrong or what's painful is fairly easy. Coming up with a solution to the problem, may sometimes be a bit more complicated. I, again, asked everybody to write down their suggested solutions to the impediments that were raised. When finished we took turns of reading the solutions and wrote down these suggestions on the board as well. This uncovered two interesting issues:
  1. Unsurprisingly - some of the problems had the same solution
  2. For some of the problems we couldn't find a solution with a good ROI (ROI = Return Of Investment: Meaning - all the solutions we found weren't worth their effort) 
This step made us focus on a smaller set of problems and their solutions. And again, everyone had their chance to suggest solutions.

Lastly - we added our suggested solutions

Step 3: Let's vote!

Lastly, I told the team that each team member gets 10 points. Those points were used for voting on the solutions that were suggested in the previous step. Each team member was allowed to distribute the points in any way (s)he wanted to. At the end, the top 3 solutions were chosen as our action items for the next sprint.

Step 4: Leave with a good taste

Before wrapping up I asked everyone to say what they thought was good during the sprint. I wanted the retrospective to end with a positive note. Also, it's important to preserve what we do well.


After the retrospective, I again did a little retrospective with myself. I tried to figure out if I managed to address the 3 problems I tried to face at the beginning. The democratic nature of the retrospective allowed everyone's voice to be heard, as I wanted, and prevented unnecessary arguments about the priorities. Also, at the end, we had to vote on some action items. Having time to think about what's really important and what we can really change, made us choose action items that were feasible. Those action items were tracked during the next sprint - I sent an email that sums up the retrospective right after this meeting. During the sprint I sent a reminder when we encountered an issue that was brought up in the retrospective. Finally, before the next retrospective I sent an email describing our decisions from the previous retrospective and whether we applied the action items or not. At the next retrospective we made sure we followed on those action items, but that's another team member's story to tell :)

A word of thanks to Gal Zellermayer, my team-leader at the time, who pushed me into thinking out of the box and conduct this retrospective.

Find me on Twitter: @AviEtzioni

More interesting posts from this blog:


  1. Very good. I liked the democracy and prioritization in your process.Thinking out of the box is something I am always looking for :)
    My small note:
    My biggest problem with the retrospective process is that I feel that most of the top-priority and common problems tend to reoccur sprint after sprint. I do see some of them on your imaginative board as well:
    * Research was needed - Have a spike.
    * PO not clear - Ask for clarifications.
    * Took too much time - Take more buffer.
    Even if we get some improvement each sprint, I do feel that we sometimes waste our time talking about the same issues over and over again.
    I would be happy to hear your comments:
    * Did you ever manage to get a big enough improvement with these big permanent problems, so they were really taken down in the priority list?
    * What so you think about trying to refrain from talking about these problems on every other retrospective meeting?
    Any other idea would be welcome :)

    1. I totally agree, and this is why it's important to track your decisions and make sure you follow them. In the "real" retrospective that I conducted we talked about much more "advanced" issues than taking more buffer or do more spike. Because we were passed this already, that was obvious to us.

      I think tracking have a constant follow-ups on your decisions is key here in order to improve and not to waste your time on the same issues over and over again.

  2. One trick we've used in our retrospectives, was to go over the decisions made on the last retrospective and see if we were getting better at it since then. If it was still a problem, we wrote it down again for the next retrospective.

    Also, we made sure that each action item would be represented in our Technical Backlog, from which we picked 2-3 items (~10-15% of the total time) every sprint. This way we allocated time to address pains raised by the team, and check them every sprint.

  3. Very nice!
    We managed a similar solution in my last company:
    First we display the summary, statistics and status of the stories and work.
    2. Every team member has time limit (yes - we use stopwatch) to say maintaining and improving things without finding a solution.
    3. We vote for what is the most impotent (every one has 2 voice/points)
    4. We talked about the most impotent things
    5. Took action items, take a photo of the board and check if we solved those items next iteration.

    1. Thanks for the feedback Itai! I liked the stopwatch, sounds very interesting and I'm about to try it sometime!

  4. Great Idea.
    I know I not from this world but i have one suggestion. After the round with the solutions is done I would add one more round inwhich evreyone will say is opinion about the solutions that were offered and after that I would make the vote. I think that in that way people can think about more upsides and downsides on each idea.

  5. Great initiative. I will definitely try some.
    My contribution is that you can have few types of retrospectives;
    1. Sprint - like you did.
    2. Retro of Retros - Where the all team are going over all the decisions/actions which were brought up in the last 5-6 retrospective and see how they were handled. It requires keeping the summary of every retro (and not just keeping the last).
    3. Problem - where you attack a single pain.

    In addition we added 2 metrics which each team member grade every retro:
    1. How do I grade myself in this sprint (more like a feeling, was it a good sprint for me? Did I manage to get fast progress)
    2. How do I grade the entire team in this sprint (How good wast the collaboration etc.)

    we found some interesting things while looking on those metrics over time identifying peaks and down times.

    1. Amit - Thanks a lot for the reply. You gave me some great ideas. Especially, I think I'll apply the "Retro on Retros" in order to make sure things happen. Also you can have a "Retro on Retro" to figure out if the team finds the retrospectives productive.

  6. Hi Avi,

    I just tried your style of Retro at my work place and it worked really well. I will write up the impact it had and its feedback this week.

    Good stuff dude.

    1. Thanks a lot Dave! Glad you enjoyed it and I'd love to hear more from your experience.

    2. Hi Avi,

      Thanks for the blog post and inspiration. You can find my views and what I learnt here

      Cheers dude

    3. Thanks a lot for the credit Dave! I only briefed through it but I will read it thoroughly soon and add my comments.

    4. Cool no problem. Yeah be interesting to see your thoughts on it.