As those of you who have attended my workshop in agile planning know, you can talk about planning in two dimensions:
- The percentage of your capacity you want to allocate to planned work.
- The planning horizon: how far ahead in time you want to allocate that capacity.
The Scrum Guide implies that a team of developers will allocate 100 percent of its capacity to planned work, divided between implementation of backlog items that the team thinks are ready for development, and refinement activities where the team makes sure that the upcoming backlog items will be ready for development in the future. The planning horizon is a sprint: a regular period of at most one month.
However: the nemesis of all planning is that pesky little thing called reality. During a four week sprint, there are many things that can (and will) happen, so many teams feel that they have to shorten that time. Sprints that are two weeks long have therefore become popular even if it is hard for many to actually deliver value in such short time.
But since reality happens even during two weeks, you might feel the need to replan in the middle of the sprint. And this is where the rolling wave can start to happen if you let it.
When mid-sprint replanning becomes a regular necessity due to circumstances beyond people's control you may ask yourself what planning horizon you should have. Of course you need to replan the rest of the sprint (the last week), but can we say something about the future? The coming sprint?
(The replanning is by the way often conducted directly after daily stand-up monday morning in the second week of a two week sprint.)
The last week we learned that it wasn't wise to allocate 100 percent of our capacity to planned work. Maybe 80 percent is more realistic, so let us do that for this week too. And what do we know about the coming sprint? Maybe we can see the need for like 50 percent? At least for the first week, we know actually very little about the second week of next sprint. Maybe we for the second week of the next sprint can think of work worth about 20 percent of our capacity.
The week passes and it is time for sprint planning for the next sprint. We have already planned for 50 percent of the first week of the next sprint, and 20 percent of the second week. Let's just add some more insights to that plan so that the first week of the sprint becomes planned to 80 percent and the second is filled to 50 percent. (And maybe also look ahead for the coming two weeks after that but not more than like 20 percent of the first week and 10 percent of the second.)
Shouldn't we aim for filling both weeks of the sprint to 80 percent at sprint planning? No, we can't predict the future anyway. Let's keep it planned to 50 percent. We know that we will have a replanning event in a week anyway.
TADAA! The rolling wave is formed. Sprint planning becomes basically an adjustment of the rolling four week forecast that we polish continuously. We can still have our reviews and retrospectives every second, third or fourth week if we want to, but we have a more flexible planning practice in place that in many cases can accommodate the natural variation of development better.