Most of our projects run on a two-week sprint cycle, so in theory we should be enjoying many four hour sprint planning sessions. Four hours of planning! How could anyone stay awake for that long? How could we justify that from a budget perspective?
Even with detailed discussion of the story needs, using poker for story pointing and detailed planning per item, two hours would push the boundaries of a useful planning session - the team will have had enough and the quality of the discussion will degrade.
Likewise, not estimating at all is very risky and most likely to run the project into difficulty sooner or later. So you definitely need an estimation technique, but finding something which suits your project can take a little effort.
Recently, we’ve chosen to use t-shirt sizing and points for estimation on a project, which is a super easy lightweight technique for getting teams to estimate their work.
When to use it:
- Fast projects with 1-2 week sprints - you simply don’t have the time for long estimation sessions so something the team members can do ‘offline’ is great.
- For teams getting started - teams can often relate much more easily to t-shirt sizes than the Fibonacci sequence used in poker cards.
- For teams who prefer to estimate in days - if you’re trying to convince your team to use an abstract method for estimation rather than days, then this is a low barrier to entry.
How it works
- The team use T-shirt sizes (e.g. XS, S, M, L, XL) to estimate and agree work on the project.
- The team should pick from the backlog something which they agree is the smallest task.
- Each story or task in the backlog is then estimated using the T-shirt scale. Team members may just be able to estimate their own work if it just applies to them, or agree with other team members on those tasks which they will collaborate on.
- At the start of a sprint, the team should estimate or amend the estimate in t-shirt sizes for a specific task or story. This can be done either in a very short planning session (i.e. 15-30 mins maximum), or do it when they have time.
- During, or at the end of the sprint, the team should update the estimate for the task. A small task may have mushroomed, likewise something which was large may have required little effort to complete.
Tracking progress
In the case of this project, we converted the t-shirt sizes to points i.e. XS - 0.5 point, S - 1 point, M - 3 points, L - 5 points, XL - 8 points). As PM, I could then use the points to review progress using relative estimation, and estimate with greater confidence (and happiness) where we would get to by the end of the project.
It’s by no means foolproof. It’s a very simple technique, and that means the level of accuracy can be off at times. But it’s way better than trying to estimate using days.
How do you and your teams estimate your work at the start of a project? Join the conversation on twitter @clearleft.
