Why aren’t Agile estimates time based ?
The short answer is that we pretty much suck at it.
But first lets be honest, estimate is just a more corporate way of saying guess.
If you were to ask a group of people to stand in a circle, and without counting sit down when they think 15 minutes had passed, you would probably find the first estimators start sitting down after 5–8 minutes and the last few start sitting down after around 20–25 minutes.
Perception of time varies greatly from one person to the next and can be influenced by a range of factors from emotional state to the time of day.
What we are quite good at though is relative sizing. I have heard it hypothesised that it goes back to a time when we needed to know which rock would most effectively smash a dinosaurs skull, but whatever the reason. If you were to show the same group of people the following diagram:
Then ask, which is bigger ‘A’ or ‘B’, naturally everyone in the group will answer‘B’. Similarly if we ask approximately how much bigger than ‘A’ is ‘B’ most people will say it is roughly ‘4′ times the size:
This is the thinking behind Agile estimation. We are trying to capitalise on what we do best, or at least better.
In order to relatively size things we need an arbitrary counting system. Some teams use T-shirt sizes: S, M, L, XL, XXL, others use colours. My personal preference is for the modified Fibonacci sequence of: 1, 3, 5, 8, 13, 20.
To apply this to the above example, if ‘A’ were to to be considered a ‘1′ then ‘B’would most likely be a ‘5′.