> Why Should We Measure Engineering Productivity? I'm still waiting for the answer. It was not addressed in the article. > [cited as valid reason not to measu

> Why Should We Measure Engineering Productivity? I'm still waiting for the answer. It was not addressed in the article. > [cited as valid reason not to measure:] You can’t afford to change the process/tools right now I'd argue that cost to change vs cost of not changing is itself often a fluctuating value. Seeing trends there and being able to make estimates/predictions about the costs of tradeoffs is valuable. Especially when trying to change a system. > The results will be used only as vanity metrics to support something you were going to do anyway Those are most metrics I've seen. At least the ones that bubble up the management chain, in my non-Googley experience. I realize it's a really twisted version of Goodhart's Law[1], but I claim metrics should shape human behavior. Any metric you can't respond to is useless. Metrics you will never respond to are useless. Metrics that potentially make you change behavior are meaningful. This was touched on in the article but the actual human behavior aspect seemed buried, to me. [1] https://en.wikipedia.org/wiki/Goodhart's_law https://en.wikipedia.org/wiki/Goodhart's_law