Backlog Grooming includes the team responsibility for ensuring that the ENTIRE Product Backlog is estimated at all times. This means that, while we will spend 70% of our time discussing, slicing, and estimating the highest priority items, we must still do “deep dives” in the backlog, ensuring that all items have an estimate on them.
Estimating large items in story points is actually EASIER than estimating smaller items, due to the fact that the large items are lacking significant detail and, therefore, the amount of detail you have to discuss and review is limited. Because of the lack of detail, your accuracy is similarly reduced. However, because there are usually a large number of items to be estimated (more than 40 is sufficient), statistically you will estimate some higher and some lower than they will actually be – meaning that the average of the “wrongness” of your estimates will actually be reasonably precise. Also, remember that you will be discussing, slicing, and re-estimating these items anyway as the project progresses. Greater precision will come later.
Given that we have already covered how to do estimations, this blog post simply seeks to provide some tips you can use to help your teams successfully do large item estimation.
Don’t spend a lot of time on each item. You only know what you have in front of you and what your Product Owner can quickly provide. Don’t solve the problem (i.e., don’t figure out how you’re going to build it). Give yourself a time limit (perhaps 5 or 10 minutes). Then estimate.
If you can, capture the information you discuss in the process of estimating the item as comments on the item. It may help later in remembering what you discussed when you did the estimation.
Compare the larger items to the backlog items you’ve already estimated. If a backlog item seems to be a little harder than your 5 story point items, maybe it’s an 8. Harder than that? How about 13?
Don’t just come up with a bunch of new story point values (e.g., 20, 45, 60, 90, etc.). Extend the point sequence you are already using to accommodate the larger items. If you are using a Fibonacci sequence, you’ve probably already used 1, 2, 3, 5, 8, and 13. The next numbers in the sequence are 21, 34, and 55. The concept of a Fibonacci or doubling sequence works — don’t break it.
If an item is REALLY big and complicated, try slicing it into two or three smaller items. This is an expensive option (because of the effort involved), however, and a last option should all attempts to estimate fail.
Like small item estimation, large item estimation takes practice. The more your team does it, the better they get. For that reason, you might want to estimate ten or so large items and then do a quick review of your estimates to see if they still make sense.
Take your time. Follow the process – it works.
Be sure to continue by reading part six!