This excerpt from the Kindle pocket guide, Practical Estimation, discusses the relative merits of the most common units used in relative estimation.
Abstract Estimation Units
When estimating the traditional way, a discussion of estimation units doesn’t happen very often. If someone asks you how long it will take to complete a task, they’re thinking in terms of calendar time. The implication is that they expect you to answer in kind.
But it doesn’t mean that you have to estimate in calendar time. Relative estimation necessarily means that you’ll be working in abstract units. Some examples of this are:
- Ideal Time - An ideal day is a work day free from interruptions and other overhead.
- T-Shirt Sizes - Some teams use Small, Medium, Large and eXtra Large to indicate the time required to complete projects.
- Points - Abstract units representing relative time to complete units of work.
When I first started practicing relative estimation, I was taught to think in terms of ideal time. If, for example, I believed that a task could be completed in 1 ideal day, and the team’s velocity was 2.5 ideal days per week (0.5 ideal days per day), then I could expect the task to take 2 days to complete. But I found this conversion between ideal days and real days confusing. The technique also had the unwanted side-effect of encouraging me to think in absolute time instead of relative time, so was really no better than estimating in real days.
The use of T-shirt sizes has a certain appeal, in that it provides a mechanism for quickly conveying high-level information from developers to managers. In my experience the drawback is that, being non-numeric, these units cannot be used to perform arithmetic, so cannot directly be used to predict schedule.
I prefer the use of abstract points for relative estimation. Point values can be used to express the relative time it will take to complete a unit of work, compared to other units of work. Early on I was encouraged to think of points as representing the “size” of a unit of work or the “effort” required to complete it. Those terms, although somewhat helpful, never seemed to describe what it was I was actually estimating, and I eventually came to think of estimating in points as estimating in terms of relative time.
You can read more on this subject in the article titled What Are We Estimating, Anyway?)
Read more about estimation in the 99-cent Kindle pocket guide, Practical Estimation.