Retrospective meeting
Retrospective structure
Any meeting deserves a good structure. A default structure for a retrospective meeting looks like:
- Set The Stage—to get everyone’s attention in the room (flaps down!)
- Gather Data—to get everyone on the same page (just the facts, no feelings)
- Generate Insights—to find out what hurts most
- Decide What To Do—to implement one single improvement item
- Close The Retrospective—to collect improvement actions for next retrospective
The Navy Seals only use three questions to review missions. Their breakthrough, and how it will radically improve your performance: AARs (After-Action Reviews) review projects, missions, training exercises or any other event you want to reflect on. There’s a world of difference between an effective AAR and a post-mortem bitch session. The magic is in the phrasing of the questions.
Here’s how to do them:
- What Went Well—Start every AAR by asking What replicable new learning did we gain from what went well?. In any post-mortem, our brains can naturally focus on what went wrong. However, it is critical to get clear on what went right so the good can be repeated.
- What Did Not Go Well—Next, move on to constructive criticism. Frame the conversation with a similar question: What replicable new learning did we gain from what did not go well? Note that the question isn’t asking what went wrong, but specifically what is the new lesson we can incorporate in our future behavior.
- New Standards—Finally, combine the insights from the first two steps, and assess the following: Drawing upon questions 1 and 2, what changes can we make to our processes to systematically improve our consistent excellence?
This reinforces that the standard is excellence, clear process is how we get there, and asks what improvements to our process move us ahead?
- Each participant should complete these questions in writing 24 hours in advance.
- AAR sessions begin with reading all of the written feedback.
- Then the meeting owner leads the discussion of each question.
With excellence as the standard:
- What went well—keep doing
- What lessons exist from what didn’t go well—stop doing
- How will we modify our processes going forward—change doing
Default retrospective
Set the stage
- Two truths and a lie.
- Reiterate the retrospective prime directive.
Gather data
- Present outcomes of previous improvement actions.
- Present standard data like:
- Average cycle or lead time.
- Average throughput.
- Predictability.
- Changes in team composition and/or availability.
- Create a timeline of the period under review; list the days of each week (Mon–Fri).
- Collect events that happened. This is neutral, objective data, e.g.
- Mon: start using the new server
- Tue: deployed five stories
- Wed: had a beer
- …
- Create three swim lanes as timeline
- Set tick marks for every day.
- Create four typical swim lanes:
- neutral :-|
- mad X-(
- sad :-(
- glad :-)
- Optional: add lanes for the other two emotions:
- afraid 8-[
- guilty ^_^;
- Collect facts & feelings: events and observations in appropriate swim lane, cluster at will.
Generate insights
- Create table with three columns:
- Good—behavior and practices you want to hone.
- Bad—behavior and practices you want to improve.
- Ugly—behavior and practices you want to stop.
Decide what to do
- Split table into top and bottom halves, thus creating six cells in total:
- Top: Me/We (within team's scope).
- Bottom half: They (beyond team's scope).
- Generate measurable actions and goals in each of the six cells.
- Order them—when will you take action on which improvement item?
- Use the Good to try and fix the Bad and Ugly.
Close the retrospective
- Help, Hinder Hypothesis; or
- ROTI.
- +/Delta
Retrospective questions
- What went well?
- What can be improved?
- What have we learned?
- What do we still not know?
- What still puzzles us?
- What wishes do we have?
- Which single experiment will we do (to speed up)?
- What did this iteration produce?
- What was the team aiming for?
- How did the result meet (or not meet) expectations?
- What’s going on elsewhere in the organization that affects the team as they go into the retrospective?
- For example, are there rumors of layoffs?
- Has there been a recent merger?
- A canceled product?
- What is the history of previous project reviews?
- What happened?
- What was the follow-up?
- What are the relationships between team members?
- How is their work interdependent?
- What are their personal connections and working relationships?
- What are team members feeling?
- What are their concerns or anxieties?
- What are they excited about?
- What kind of outcome will achieve value for the time invested— both for the retrospective sponsor and the team?
- How has the team worked with facilitators before?
Keep your retros fresh
- Also conduct retrospective meetings at other times than in between sprints.
- Consider to ‘good bad ugly’ them, and physically crushing the ‘bad’ and ‘ugly’ after having collected them, and then ‘perfection game’ the ‘good’.
- tip top each other, just like a temperature reading. Top identifies something you value in the other. Tip is a request—petition, solicitation, prayer, desire—for specific behavior of the other.
- Turn the focus outward and ask yourself, “What can we give back to our environment?”
Alternative retrospective Format
Pick two of these three key focal points in mind for every retrospective:
- speed;
- fun; and
- quality.
Sources
- Agile Retrospectives—Making Good Teams Great by Esther Derby, Diana Larsen
- Signal v. Noise » Claire Lew » The 9 questions that uncover the most surprising insights from employees
- InfoQ » Shane Hastie » Linda Rising on Continuous Retrospectives
- Agile Cafe » Omar Bermudez » Story cubes: Build your story with cubes in the next retrospective
- Mastering the Obvious » Ellen Grove » Rollin’ Rollin’ Rollin': Using Story Cubes to jazz up team retrospectives
- InfoQ » Ben Linders » Adding Purpose and Hypotheses to Agile Retrospectives
- InfoQ » Ben Linders » Having Actions Done from Retrospectives
- Target Process » Michael Dubakov » Development practice: Retrospectives in Kanban
- Seth’s Blog » Seth Godin » Self cleaning
- InfoQ » Luis Gonçalves, Ben Linders » Getting Value out of Agile Retrospectives - A Toolbox of Retrospective Exercises
- InfoQ » Rui Miguel Ferreira » The Power of Anonymous Retrospectives ← must have its own anonymous retrospective
- LeanKit » Chris Hefley » How to Run Effective Standups and Retrospectives
- Seth’s Blog » Seth Godin » What’s next?