Let me have a bit of time to explain myself. But the subject came up too often in our peer coaching sessions this week. I’m going to push the cork first: “grooming” (refinement) is a mistake. Like the “definition of ready”. What they’ve become is window dressing. Now I’m going to soften my words and explain the perspective I’d like to share as widely as possible.
Grooming was born to address the situation: right in the middle of sprint planning the product owner came with a user story and boom: poorly broken down, or too big, or terrible dependencies being discovered, etc. And so it’s chaos. Apart from the fact that chaos is often a source of creativity, I can understand that it’s exhausting and while it may be creative, it’s not necessarily productive. In short, grooming was born to anticipate these moments of chaos and make sprint planning flow better. Rightfully so.
Problem
Grooming sessions have become obese. We do two or three per iteration (of two weeks). They last two hours. And that’s a real problem. And it’s also the signal that you’re not approaching subjects with an agile approach, for me it’s a problem, but these are my convictions. Why a problem? Why not agile?
A conversation not necessarily shared
In grooming sessions many people have gotten into the habit of having conversations in sub-groups, not with everyone, ultimately it’s about anticipating to avoid the chaos of sprint planning, two or three people answer the questions. But when sprint planning arrives everything is settled. The sprint planning becomes a hollow rubber stamp. And the true great group conversation of the iteration has disappeared. We’ve lost the interactions that are so important.
Too much conversation not enough action
Opposite scenario everyone is there, and we talk for hours. In a sense: we think too much before acting. And a large part of the answers, as always, will come through action. We anticipate too much. The duration of your conversations should be relative to the duration of your iterations and proportional to your capacity for delivery. If you have so many conversations, it’s because you anticipate too much, and you’re trying to answer too much in advance. Precisely what an agile approach tries to avoid.
The right conversation at the right time: sprint planning
Ultimately you should just deal with what’s coming soon, and you shouldn’t answer it definitively in a sub-group, but have enough information for it to be dealt with serenely in the sprint planning which would regain its full meaning. The sprint planning is the start of the iteration. It also makes sense because the review and retrospective took place before, and normally, they brought new information to the product owner and the team.
Step by step
People ask me (during these peer coaching sessions): but what about big dependencies, big subjects? As always big subjects are treated step by step by priority of importance, of value. If you need some kind of big initial clearing that’s perhaps precisely the occasion for these quarterly planning (or PI planning), to have a quarterly approach say or at the start for two or three days you share and deal with these “big” questions.
Signals to observe
- Your sprint planning lasts several hours (2 to 3h let’s say) and that’s where the debates take place. It’s a good signal.
- The grooming conversations are only anticipations to make the conversation possible. It’s a good signal.
The fundamental problem with obese grooming
Grooming sessions that are too frequent, too long, are there to justify dependencies (like the pi planning board) instead of resolving them. Dependencies must be limited. Obese grooming sessions sanction them. Bad grooming serves to maintain these dependencies. Grooming grows fat because they exist and we don’t seek to resolve them. This is exactly what must be avoided.
My Scrum experience
I encourage you to question the people I’ve been able to work with in this context.
Conclusion
Yesterday morning I wondered, confronted with the recurring question of grooming and then also suddenly confronted with this heresy about the scrummaster in the scrum guide (next article), I wondered: am I condemned to repeat the same things until the end of my time? And then, I wondered: is it a problem? Maybe not. I’m ready to try to explain my point of view ad vitam aeternam it’s perhaps my role, it perhaps lacks humility.