I was Wrong about Re-Estimation

I’m just starting to learn and implement Scrum without yet having completed the CSM training. I’ve years of reading about Agile Software development including the original Scrum Book. So I thought I had a solid understanding. However putting theory into practice always reveals questions that I’d never considered.

Recently on the Scrum List I asked whether people were following the book on re-estimation:

In the first scrum book Ken talks about re-estimating the tasks daily within a sprint. Is that how people are working?

We’re working our way through our first sprint and are finding a number of small tasks that we didn’t think of during the sprint planning session. Do we generate new estimates for these? Do we do a daily update of the tasks we did think of?

My original leaning was not to do this. My thinking: During every planning session there will be some tasks that we miss and others that we estimate poorly. So instead of trying to update the estimates constantly over the course of our first few sprints I expected we would learn roughly how many estimated hours we could complete in a sprint.

When it comes to updating the estimates constantly – it just seems like additional overhead with only some value. What do you do?

Wow, I couldn’t have been more surprised at the response. Most teams take a few minutes to recheck their estimates and add newly discovered during the daily stand-up. Newly discovered tasks are added to the Sprint backlog and the burndown chart is updated. This gives everyone a birds eye view of how the sprint is progressing. A little bit more work for the ScrumMaster == a lot more value for team (and management).

Thanks to Jeff Hienen, Leandro Saad, Thomas Stagl, Rachel Davies and Alexey Krivitsky for correcting a beginners mistake.