Has Scott Ambler lost the plot on measuring agile performance?
I know this article was written back in October, but I can’t believe that nobody has pointed out its flaws. No Scott, measuring “acceleration” is not the same as measuring delivery of value!
Scott Ambler, an author who I generally have a lot of time for seems to have completely lost it in an article at IBM developerworks. In this article he goes to some lengths to justify measuring teams on “acceleration” (in effect, measuring the proportional increase in story points processed per iteration) and thereby assessing their value. While this may have some merit, the article insists on contrasting it against other methods of attempting to measure actual productivity.
Unfortunately such a comparison does not make sense.
- A high-accelaration team may be worse over time than a low-acceleration team, if their starting productivity is less.
- A high-accelaration team may be worse over time than a low-acceleration team, if their maximum productivity is less.
- Ambler’s analysis assumes that the entire surrounding process is constant, including task estimation. This is very unlikely
Fundamentally Acceleration and speed are quite different things. I can wait at the lights on my pushbike and burn off nearby cars, because of my higher acceleration, but fifty feet later they will overtake me and carry much more stuff a much further distance.