I read her other post from 2004 and i can't help but think you're throwing the entire thing out, baby and the bathwater and all.
I don't think you should make the metrics the end all be all (Goodhart's law), but that graph is certainly helpful if you can figure out who might be doing literally NOTHING for hours on end vs. someone who is productive at some level. All trying to "figure it out" at that point as a manager is just trying to cut through someone's bullshit when the data is right there.
Maybe it's more like 90% of the cases those metrics shouldn't be used to measure anything, but they can certainly point out "smoke" where there might be someone struggling and then you can be an actual manager and figure it out case by case.
Yea because churning out or changing code is the only imaginable productive thing a software engineer can do ? what about planning, helping others out, researching all this stuff is super difficult to measure. It all depends on the work, the team and the size of the company i guess.
I totally agree with you, but there've been employees I've managed in the past who loved to go off and get distracted with anything they judged "useful", often at the expense of their actual work. That's something to be managed rather than measured by metrics.
did i say that? that's why i said it's an indicator, one of many. if you are producing zero code vs your peers and your job is to program it doesn't mean you are unproductive, but at least someone can talk to you about it and clarify vs. just guessing with zero data.
i don't understand how this is punishing the good people... i think everyone here has some ptsd with terrible managers or others micromanaging their work. data a good way to look back and ask the questions that might need to be asked, but to not use the data as the final criteria for anything (which is where i think most lazy managers end up).
do companies plan their strategy with zero data? i find it hilarious we devs somehow think we are a special group that can't be measured at all, so just don't bother and let us be. at the same time we don't want managers up in our business all the time either. just because the measurement isn't perfect doesn't mean to not measure at all.
i mean i don't considere issues closed or tickets open to be "looking over my shoulder", but i guess we disagree there. these are some metrics that need to be looked at at some point to determine some strategy about the future or reflect on the past.
"trust" to me is knowing that everyone that has access to these metrics are making decisions in good faith. it's knowing that your boss isn't sitting there watching it every hour or using it as some simple metric to axe you. blame the people not the tools.
I don't think you should make the metrics the end all be all (Goodhart's law), but that graph is certainly helpful if you can figure out who might be doing literally NOTHING for hours on end vs. someone who is productive at some level. All trying to "figure it out" at that point as a manager is just trying to cut through someone's bullshit when the data is right there.
Maybe it's more like 90% of the cases those metrics shouldn't be used to measure anything, but they can certainly point out "smoke" where there might be someone struggling and then you can be an actual manager and figure it out case by case.