Estimation

Estimation is predicting how much effort is needed to complete the work ahead. Estimates are often wrong but the process of coming up with them is useful. Estimation can be viewed as a knowledge gathering exercise in which you uncover information as you remove uncertainty.

The benefits of estimating are:

• It gives us a chance to start managing the uncertainty that lies ahead. It allows one to discover misconceptions, inconsistencies, and areas that need further investigation early on.

• It provides chance to start digging into the work that lies ahead in order to understand it, which creates a shared understanding of work in the team.

Despite the benefits of estimating it is important to understand that it is not useful to trying to make the estimate accurate or spend days investigating before providing the accurate estimate. Estimate is an estimate and not actual so quick investigation and discussion is enough to provide the estimate with the goal of uncovering uncertainties, risks and dependencies.

There are multiple tools used for estimation:

• Story pointing is relative estimate that compare size of work items rather than giving the exact estimate in hours or day per work item. Usually Fibonacci series are used for story point as the number goes the uncertainty and risk increases. If a story is 2 story pointer and another story is double the size of the story, the story points will be 5.
• T-shirt sizing is also a relative estimation technique but it uses S,M,L instead of using numbers.

The discussion is the goal, of the estimation exercise, as discussion is where learning takes place, new ideas are discovered and old truths which we carry around get challenged and may be discarded, scope of work is clarified. Splitting up work into smaller increments also outcome of estimation process. Coming up with no estimate because of too many unknowns is valid and useful result too.

When estimating include people with different skill level and roles, as it will help see different point of views. If the variation in estimate is large ask questions like:

• What makes this story 5 for you?
• What leads you to believe this is a 8?
• Why do you think this is similar to the work in X (where X is a completed story that had a 8 point value)

Goldilocks estimation technique:

Goldilocks estimation technique suggests to change the work items to right size instead of providing the estimate. The work items can be changed to right size by either merging small work items or splitting large work items. The out come is a backlog of average sized work items, which makes the forecasting easier.

To do goldilocks estimation exercise, at first split the work items in three stacks: too small, just right and too big. Then for each work item have team vote on which stack to place it on. After this have team split work items in too big stack to make it right sized and merge the items on too small pile to make them just right sized.

Are you using any other estimation technique? Please share.

Go back to the Essentials page.

Reference: Kanban In Action by Marcus Hammarberg and Joakim Sunden