The title is not as contradictory as it may seem: Agile and scrum stress the need to be flexible, to change and adapt. But let’s not forget that there’s also merit in the steady, the regular, the predictable. Humans are creatures of habit and I’ve found that the chances of successfully changing behavior increase if you create a reoccurring time for the behavior. Steady dates create a consistency and rhythm that help to trigger and internalize new behavior patterns.
life work through regularity:
- Each sprint we have several timeboxes, so-called “PO slots” to talk about details of upcoming stories. When we started scrum those slots were at random times. The developers kept missing slots they needed to attend.
Soon afterwards the POs switched to a schedule of 3 times per sprint (Fri, Tues, Fri at 12:00) and the developers usually plan PO slot attendance during standups. They rarely forget it and if they do, they remind each other.
- We had trouble fitting bug fixing into our scrum implementation (when the bugs were unrelated to stories in the running sprint). Bugs had no clear priority and sometimes stayed unfixed indefinitely. Then we introduced a 2h-slot per week during which all developers fix bugs and that works pretty well.
- Have someone take care of repeating tasks on a per sprint basis
- There’s a consistent lack of developer feedback and votes on stories, because most teams delay that to the very end of the sprint and examine the upcoming stories in one long ad-hoc meeting.
Only one team votes consistently, because they broke the task up into little pieces: They examine the stories every day for 15 minutes after the standup. Added bonus: They know what stories they want to work on and are able to clarify questions well in advance of the planning meeting.
Another experience: The chances of actually doing $newBehavior are higher, if it’s a regular appointment and you cancel when it’s not needed, than the other way around. If you need to explicitly make the appointment whenever it is necessary, it won’t work half as well. I.e. opt-out is successfullier than opt-in 🙂