Good metrics are the ultimate business tease, offering tremendous promise, making it sound easy, and then falling apart at the last moment. I think that humans become seduced by the logic and stability of math’s reputation, forgetting the fact that numbers, like humans, can also be irrational, chaotic, misleading and untrustworthy. If left to our own mental devices, human managers kvetch over most decisions, especially those with critical financial impact, like when to accelerate which engineer’s work, which set of customers to design features for, and how much to spend to ensure product quality. If you offer these managers a measurement that promises to provide the objective red or green light, you can probably get their attention pretty easily. Unfortunately, plug and play metrics are too often like pyrite, shiny but worthless.
As is the case with all best practices, context is king. Relativity rules the roost, and any metric you adopt should be vetted for your specific environment. Many people gain this enlightenment by learning from failure.
If what I’ve described speaks to you and you have a need for a sound approach to measuring product development activity, read my latest column in Time Compression, “Metrics Matter…”