Agile Software Development

Stop Re-estimating Your Stories for Every Iteration

For some, it is a common practice to estimate stories again at the start of each iteration to update accuracy. Adrian Wible hates this idea. In an article for AgileConnection, he lays out an argument for why this practice is not only a waste of time, but actually damages the final results.

Where the Numbers Diverge

When a story is being broken into tasks to be inserted into an iteration, that is when the team has the most clarity into it. In Wible’s view, this causes estimates to likely grow. The other component to this though is that teams want to meet commitments, and they may unconsciously inflate estimates so that they fit into a comfortable commitment level. To illustrate this concern, Wible gives an example of an agile team who consistently delivers a fat 40 points of velocity, and maintains a perfectly prioritized backlog with nice and tidy stories ready to grab:

…the team begins peeling off those stories until the iteration is full.

The team re-estimates the stories as they peel them off, of course. When they get to 40 points, they declare the iteration planning done.

You revisit the product backlog and see that the aggregate estimates for those stories was actually 30. You might applaud the team for gaining more accuracy—after all, the original estimates were off by more than 30 percent. In fact, if you go back over the past ten iterations, you see this is a consistent pattern. Thank goodness for re-estimation.

Now here comes the real problem. Wible asks us to imagine that the business sees a remaining 120 points’ worth of work, and since the team always executes 40 points, they assume three iterations remain. It will ultimately end up taking four iterations though, because the 120 points and the team’s 40 points exist according to different sorts of calculation. In this case, re-estimation has distorted and confused business productivity.

Wible understands that such severe discrepancies will not always happen, and that most re-estimations will probably break even in their assumptions over time. If that is the case though, and re-estimation does not ultimately impact the bottom line, then why is it being done? Wible acknowledges there are indeed times where re-estimation has a place, but they are few and far between.

You can read the original article here:

Show More