Most teams judge a retrospective by the energy of the conversation and the number of action items on the board at the end. Full board, lively discussion, great retro.
We’ve concluded the opposite: a board full of good ideas is a warning sign.
A recent conversation with a pair of dev leads revealed an example of this. “We bring a lot of good things to the surface and come up with a lot of action items, but we never prioritize them. Nothing’s actually happening.”
Almost every team we meet has a graveyard of retro action items somewhere, whether it’s a Slack channel or a “continuous improvement” backlog nobody looks at.
The usual advice here is about follow-through: assign owners, start each retro by reviewing last retro’s items, put up a Kaizen board. Those are good moves if you have the right action ideas and they’re still getting stuck. But often, the pile of stalled action items is just a symptom of the conversations that produced it.
We call this the suggestion box retrospective. “What went well? What didn’t? What should we try?” It can generate a lively conversation and a full board: process tweaks, a tool someone’s excited about, several things people outside the team really ought to fix. A grab bag across a dozen different issues.
There’s no good way to prioritize a grab bag, because nothing on it is connected to what’s actually limiting the team. The conversation jumped to action too quickly. And your team can feel that, even if nobody says it. That’s why nothing gets prioritized: none of the ideas are obviously compelling enough to jump ahead of product backlog items.
A retro structure that finds the right action item
As we wrote in You Only Need to Change One Thing, every system has one constraint limiting its throughput at any given time. Improving anything else is at best wasted effort. So before your team brainstorms a single idea, back up and figure out what’s actually constraining the impact of your work and the ease and flow of doing it.
This is what our favorite meeting shape, the ORID structure from The Art of Focused Conversation, does when you aim it at a retrospective. Each phase has a specific job in the hunt for the constraint:
Observation. Collect data about what happened, anything that might tell you something about the constraints on your impact and flow. Bring artifacts, not just memories: metrics, the sprint timeline, charts from your tracking tool.
Reflection. Collect a second kind of data: people’s intuition and motivation. Motivation works as an early indicator here because mastery and purpose correlate nicely with impact and flow. If the team is feeling demotivated, it’s often because the work is harder or less impactful than it ought to be. Treat that feeling as data about the system.
Interpretation. Now make sense of it all. What’s increasing our impact and flow? What’s decreasing them? Keep digging until you’re down to one likely constraint, or a small set of candidates.
Decision. Only now do you brainstorm ideas, and only ideas aimed at the constraint. Then craft the winner into an experiment and decide where it goes in your sprint plan, so it gets real capacity instead of a Slack reminder.
What this looks like
One of our favorite examples comes from a team having exactly those lively suggestion-box retros: lots of things they could improve, and often lots of things people outside the team needed to fix. None of it improved their flow of value.
When they switched to ORID retrospectives, one of their Observation inputs was a set of cumulative flow diagrams from the last several sprints. The diagrams made the constraint hard to miss: the team was starting pretty much everything on day one and finishing very little by the end of the sprint. Then they’d carry that work over, pull in more, and start all of it on day one again.
They were working on everything at the same time. It didn’t matter how much they fine-tuned other parts of the system. So they ran one experiment: lower work in progress. Their throughput went up in a big way.
One team, one constraint, one experiment. That’s the goal: a retro that produces less and changes more.
Free ORID facilitation guide
Want help facilitating this kind of retrospective? Download our ORID facilitation guide for free (excerpted from our online course, Facilitating Effective Retrospectives) by filling out this short form:
Download ORID Facilitation Guide
For a bigger-picture version of this format, see Our Favorite End-of-Year Retro Format.
Last updated