Retrospectives

The output is a change, not a discussion. A retrospective that produces observations and no action is a meeting about feelings, and people correctly stop attending.

Three questions, fifteen minutes: what worked, what did not, what will I do differently next iteration? Pick one change. One that actually happens beats five that were listed.

Look at facts, not impressions. How long did things actually take against the estimate? What broke after it was called done? Where did the time go that you did not expect?

Working alone this is easier in one way and harder in another: nobody disagrees with you, and nobody challenges you either. Writing it down is what makes it real, because otherwise you will remember the iteration as going roughly fine.