Metrics in an Agile World

Recorded at:

If there was a point I didn’t get it! The evil side of metrics is neither new nor specific to software development. For instance, measuring students’ performance through exams can make some of them study towards assessment not learning. Does that mean schools should stop trying to measure student’s performance?

The title led me to believe that the presentation is about the metrics that fit well in agile mindset and how to use them. Did I expect too much?

You did not expect too much, and I think it's a fair criticism that we didn't spend as much time on the possible solutions to the metrics dysfunctions as we would have liked. Jim summarized the approach nicely, but we would have enjoyed examining many of the metrics that the audience brought up, and looking for ways to find alternatives, or to use the "measure up" approach to mitigate dysfunction.

Moving performance metrics up to the level where individuals (or the teams measured) cannot directly alter the measure is the basic approach. Anonymous reporting, and aggregating the values, prevents that dysfunction, and turns a motivational metric into informational. It can still be used to measure performance, but it's at a higher level, and resembles something that we would really prefer to optimize, rather than setting up an opportunity for local optimization (gaming the system).

We're still exploring these techniques, and I'm hoping to shift the balance of the talk from so many examples of dysfunction into examples of practical techniques to remove the dysfunction.