I sometimes wonder if all the time spent inputing and tracking various not-naturally-generated "metrics" is worth it. Manager and executive hours analyzing this stuff and devising systems for it and arguing over which systems to use and crafting mandates and plans, can be extremely expensive. Expensive ICs like programmers can easily lose 10-20% of their time to task-tracking and time-tracking and shepherding various things along that only exist for better tracking. Like, if we just didn't do those things at all... would productivity go down? Or up?

The cost of doing all this measuring and evaluating and preparing-to-measure certainly isn't being measured, though, so we'll never know.

Yes, and often the management instituting such granular tracking systems are also the people who want to go outside it for their own "quick asks".

It's entirely possible to construct beautiful metrics, amazing velocity charts, etc and have a terrible product with miserable users. Many such cases.

Some of this metrics stuff is like "expressed vs revealed preferences". Users who complain about your product and demand new features more quickly don't want a beautiful burndown chart. They want you to improve your product. You may actually need to slowdown to speedup. It's likely you need to do some R&D, and lots of experiments that will look bad on metrics.

You see this a lot when you quantize teams into little buckets like operations vs new features vs customer success vs core engineering teams. It's possible for a team run by a bad actor to have nice metrics while being a net drag on the other parts of the org.