It's difficult, if not impossible, to measure the impact of individual SWEs. The challenge of measuring the impact of managers in charge of large quantities of SWEs is surely an order of magnitude more difficult.
I imagine this is compounded by the fact that anybody who has risen to control of a fiefdom at a prestigious place at Google is surely highly intelligent, driven, and is either highly competent or at least highly adept at projecting competence. I can't imagine spectacularly obvious dunces rise to such heights at Google.
To be clear... I'm a million miles from Google and have zero inside knowledge, just speculating
It's not hard, it's just that silicon valley is obsessed with quantitative metrics, which are fucking stupid.
Just make a judgement call. It might be completely wrong a small % of the time, which is better than being mostly wrong all of the time.
What's so tough about this though? Someone is responsible for this shit show, make them accountable. In the abstract you may be right but this isn't abstract