Comment by DanielHB

I think that a good portion of the problem is that there are groups involved in metrics:

1) People setting the metrics

2) People implementing/calculating the metrics

3) People working on improving the metrics (ie product work)

2 is specially complicated for a lot of software products because it can some times be really hard to measure and can be tweaked/manipulated. For example, the MAU twitter figures from the buyout that Musk keeps complaining about, or Blizzard constantly switching their MAU definition.

Often 2 and 3 are the same people and 1 is almost always upper management. I argue that 1 and 2 should be a single group of people (that doesn't work on the product at all) and not directly subject to upper management and not tracked by the same metrics they implement (or tracked by any metrics at all).