Magic Estimation and the right comparison stories
I am in a Scrum Team and we are working with Azure Devops for our Taskboard and Backlog. Now we have found the issue in the Retrospective, that we need new comparison stories for estimation because the estimation did not work well last sprint. The Scrum Master is on a long vaccation and it was assigned to me that I find new right comparison stories for the Magic Estimation, which we want to test.
So there are two issues here: I don't know how to find the right comparison stories for the magic estimation game. And I have never done a magic estimation. So I have absolutely no clue how to start on this task. Maybe you can help me or give me a hint how to find the right comparison stories and how to execute this magic estimation with the team.
The first thing that I see is that it seems like you are putting a lot of effort into the estimation efforts of the team. The time when an estimate is valuable is extremely narrow, so whenever a team is putting a lot of effort into defining or executing different strategies or methods for estimation, that strikes me as potentially wasteful. Why is your team estimating? What do you hope to achieve by estimating the work?
Once you know why you're estimating, the right methods may reveal themselves. If you do need to calibrate your estimates, I'd suggest that it's not an individual person's work. The whole team should be involved in this calibration activity. There are several different ways, but it is important to pick a piece of work that everyone is familiar with, agree on its size, and then everything else becomes relative to that piece of work. Some teams prefer to take a very small piece of work and make everything proportionally larger. Other teams prefer to take a piece of work that is in the middle of what they would accept. It may also be helpful to have a few examples of work of different sizes.
it was assigned to me that I find new right comparison stories for the Magic Estimation, which we want to test.
Pick any one at random and call it a medium, or perhaps a 5 if that's a medium size. The Developers then group the other stories higher or lower, or the same, in relation to this and each other.
You are likely to see a normal distribution or bell curve emerge after about 15 or 20 stories. Recalibrate with the medium at the centre of the curve. It doesn't therefore matter which you picked first.
In addition to what the others have mentioned, get all the Developers and Product Owner in a room (or together on videoconference and use something like a Mural or Miro board). Grab some stickies and place them in columns, labeled by the relative scale being used (e.g. 1, 2, 3, 5, 8, 13 or XS, SM, M, L, XL). Draw some dividers between columns.
Then write the Product Backlog items that need estimates on stickies. Have the Developers have a conversation about each item (and ask the PO questions), then they agree on the relative size, and place those stickies in the appropriate column.
Then everyone can look at the Product Backlog items to make sure there is affinity, and have a discussion if any Product Backlog item is in the wrong column.
All the best,
thx to all for your answers I will try and hopefully everyone will be satisfied and the work will be completed and successfull. :)