Agile estimating 2/4: Absolute versus relative estimates

4 Dec 2012 Erik Talboom Agile & ScrumLess is MoreWork is Play

When you think about it, people always ask for absolute estimates. How long is it going to take to fix my car, when can I have it? And they are not happy when you say between 1-4 hours or 2-3 weeks. They need one absolute number, and some don’t even care if it’s right or wrong. They just need that number or date or timing. But is there such a thing as absolute measurements? Let alone absolute estimates? If you come to think about it, what we call absolute units of time (or size, length, mass for that matter) are just conventions. So every time we measure the length of our driveway, we compare it to a meter. Why is a meter a meter? Because we all agreed that we would call that length a meter. So in a way we are always doing relative measurement.

Then wouldn’t it be a good idea to do the same for estimates? An estimate is nothing more than a well educated guess. We use all the knowledge and experience at hand to make a guess about the amount of time it is going to take. So instead of looking at every new work item separately, why not compare it to previously finished work items? It’s easier for humans to relate to similar items than to guess the actual size of things anyway. Our brain is not capable of doing absolute estimates; we always put that new thing that we need to estimate in relationship to things we already know. This is how the human psyche works. And this is something that we use heavily in agile methodologies. Not only because of the fact that it’s a lot easier for us, but also because of the relationship between effort and accuracy you can see in the graph below.

Estimating the size of a work item does not help in getting it done. On the other hand the discussions we have while trying to estimate it, are helping us to understand this new piece of work better. And therefore also contribute to actually doing to the work. But it’s still just talking, preparation, nothing more. So we want to minimize the amount spent on this work item that is not really getting it done. And if we look at the graph below, we can see that there is a point in time where the effort we put into making our estimate better, discussing more details about the piece of work, will only have a marginal impact on the accuracy of the estimate. So you’ve got to ask yourself, is it still worth the trouble?

When we have historical data about similar work items and the size of those, it can make our lives a lot easier. We can just triangulate our new piece of work: is it closer to this really small thing? Or is it more like this normal sized item? Or is it really huge like that one piece of work we finished last month? Doing relative estimates will not only reduce the amount of time spent on estimating work, it will also heavily increase the accuracy of the estimates.