I'd say most good industries don't have cold hard metrics in place, and the companies with more metrics within industries are generally the worse ones. There are engineering firms that have heavily metrics-based evaluations of engineers (e.g. bug-fix counts being directly tied to bonuses), but I don't think I've seen them being called out as good examples. The better companies usually have engineering managers who are actually in touch with what their team's doing, and can give accurate subjective judgments of who on their team is contributing what and what the relative value of their team members is. It's the companies/managers who don't have any idea what's going on in their teams who fall back on metrics to try to answer those questions.

At the simplest level: if the company doesn't meet its revenue and profit targets, I don't get a bonus.

Teachers aren't even evaluated at the level of "more exams passed than last year" or overall results relative to comparable schools.

We've seen problems with compensation linked to short-term company performance. Test scores and most metrics like you're proposing have similar problems. Nobody cares about test scores; we care about long-term success of students into adulthood, and it seems that we have no idea how to predict that.

As the article illuminates, we're not even sure which aspect of teaching is important for it.