In my last email, I wrote about how to change your mind for a recorded decision.
But how do you know when you should?
I wouldn't leave that to someone remembering to question it. And I wouldn't set any sort of schedule to review everything periodically either. You're already doing work, so let that work be the guide for when to review anything.
That being said, it's a good idea to write down some triggers when you're recording each decision. This is where you can name what could make you reconsider things.
Going back to our chart example, suppose we chose to show a short summary beside a chart, with individual entries available underneath. We could record these as review triggers:
- Testing shows that people misunderstand the summary or can't find the entries
- The chart starts showing a longer period and the summary hides important changes
- We introduce incomplete data, which the original approach didn't account for
- Testing with user with disabilities shows they can't use the chart as is
Each trigger points to something anyone can notice during regular work. This is where it pays off to be specific, because "review when needed" doesn't cut it.
I like to write my review triggers around the assumptions that mattered when I made the decision. If the approach depended on a component behaving a certain way, a change to that behaviour will cause me to review the decision as well. If it depends on people completing a particular task and I later find out that they can't, I need to review the record.
All this doesn't mean I have to reverse the decision. It just means I should check whether what I thought back then still holds true now. I might keep the approach, but amend its scope.
The biggest problem with review triggers that I noticed is that I try to predict every future problem, so that section tends to get quite long. I always have to catch myself before doing that and try to start with the few changes most likely to affect it.
A review trigger is useful if it gives people permission to question the decision and a concrete reason to do so.