Hi! We’re Alan Dayley and Tom Looy, two of the agile coaches at BigVisible. On November 30th we participated in a live debate on agile estimation to discuss the following question: “Is sizing user stories using story points a form of waste for XP, Scrum, or Kanban teams?”
But before we start the debate we want it be clear that when we talk about agile estimation, we’re talking about sizing user stories, not estimating user stories. We are both in firm agreement that estimating user stories is a bad idea. Here’s why.
When we estimate something we give it a value within the context of some sort of unit. For example, if we were estimate the number of jelly beans in a jar the units are jelly beans. When we estimate how long it will take to get something done, the unit is some form of duration, like days or hours.
When we size something, on the other hand, we are comparing the relative size of one thing to another. The units are nearly irrelevant. For example, if we have one jar of 500 jelly beans and a second with 1000 jelly beans we would be able to see at a glance that the second jar is about twice the size of the first. We don’t need to reference jelly beans in our answer. All we care about is the relative size of one to the other.
So why are we making a big deal about the difference between estimating and sizing? Well, humans are wired pretty well for sizing things relative to one another (like the number of jelly beans in a jar as compared to another) but are not wired well at alll for estimating how exact measurements like exactly how many beans are in the jar, or something even harder, like how long it would take us to eat 500 jelly beans. (Tom claims he could eat all 500 jelly beans in about 15 minutes, eating handfuls at a time. Alan prefers to savor each flavor individually and make the experience last for days.)
Duration is particularly difficult to gauge because we often are unaware of details that would heavily influence the estimate—we assume a best case scenario; we don’t take into consideration interruptions; and we are unable to take into consideration dependencies on others and their future availability.
Another issue both of us have with providing duration-based estimates, other than our inability to do it well enough to be of value, is that specific estimates (500 jelly beans, 15 minutes to eat them) are too easily interpreted as a commitment. We too quickly forget that our estimate was a best guess at the time and we don’t make accommodation for refining our estimates as we learn more. We are all to often held accountable to our initial estimate, which lets face it, are only a guess.
So, neither of us see much value in estimating user stories, but what about sizing them? Is there value in sizing user stories using story points, a unitless relative sizing measure? One of us believes so and the other doesn’t.
Take a peek at the debate!