It's impossible to know whether or not 6% includes people who actually are high performers without knowing the internals within Google. That's why I said "doesn't seem crazy unreasonable" not "is definitely not unreasonable".

It's all relative anyways, and it depends on what you're looking for. The bottom 6% of NBA players could be considered "low performers" in that context even if they would be considered "high performers".

The problem with this methodology is you could have some teams with lots of high performers and some teams with lots of low performers, and the teams with high performers could be forced to mark people as low performers who are actually better than people on other teams.

The "right" way to do that would be to say "here is what we assess to be low performance" and then work backward and figure out what percentage of people fit within that category. However, this can be really difficult as measuring performance is very qualitative and by definition relative.

Because we, as outsiders, have very imperfect information here, it's very difficult to make an assessment of "is this a terrible policy" without knowing what they plan to do with the low performers and how they're assessing what is a low performer. I'm not even sure if it's an "each department cuts 6%" or 6% is just an approximation of what their new methodology will produce.