I think a lot of people really struggle with metrics and evaluation. We've gotten used to algorithms and mathematical expressions but I don't think many have internalized meaning and the actual information that they convey. All metrics are guides, not absolute objective signals. People hear Goodhart's Law and think about how metrics can be abused, but the real lesson here is that metrics are only guides. The abuse only happens because they become absolute measures rather than only a part.
It may take one person all day to produce 50 lines of code and another 500, but we can't differentiate these people by the metric of time, number of lines, nor the number of commits. Even bug and error metrics don't complete the story. Those 50 lines could be more critical, more complex, and have higher impact.
Realistically, we need to embrace the chaos. Everything is fuzzy and messy, and that's actually okay. That's how life and the universe actually are. Metrics are good guides to this chaos and help us navigate these complex landscapes, but they all have a limitations, which if we aren't aware of can lead to disaster. You need to be aware of the uncertainty and limitations of your metrics. I see a lot of HN/STEM like people argue passionately for meritocracies, but have yet to see one of these people account for the limitations of the measures they so passionately advocate for. The irony of this being that this takes us further away from that meritocratic goal and actually increases the noise in the system. To make a system more meritocratic you have to embrace the noise in the system.
It may take one person all day to produce 50 lines of code and another 500, but we can't differentiate these people by the metric of time, number of lines, nor the number of commits. Even bug and error metrics don't complete the story. Those 50 lines could be more critical, more complex, and have higher impact.
Realistically, we need to embrace the chaos. Everything is fuzzy and messy, and that's actually okay. That's how life and the universe actually are. Metrics are good guides to this chaos and help us navigate these complex landscapes, but they all have a limitations, which if we aren't aware of can lead to disaster. You need to be aware of the uncertainty and limitations of your metrics. I see a lot of HN/STEM like people argue passionately for meritocracies, but have yet to see one of these people account for the limitations of the measures they so passionately advocate for. The irony of this being that this takes us further away from that meritocratic goal and actually increases the noise in the system. To make a system more meritocratic you have to embrace the noise in the system.