There is a significant difference between "stack ranking" and "forced distribution rating scale." Most companies use the latter, the former is much more rare.
The practice at Amazon was like Microsoft in function. %20 were "underperforming" no matter how well they did. It was rolled up from team, to group, level, etc.
Politics was a huge factor. If a team worked really hard during the quarter and got a lot done, but hadn't shipped their feature yet, they risked all bering in the "underperforming" category, simply to allow a peer team to have more "over performers" to reward them-- for actually delivering or for sufficiently sucking off the appropriate boss (or not getting knives in your back.)
There were a lot of (metaphorical) blowjobs and knives in the back in that organization.
Which is what you have to have when ALL of the management is non-technical.
You can't have a meritocracy when the people assigning ratings are incapable of judging merit.
Politics was a huge factor. If a team worked really hard during the quarter and got a lot done, but hadn't shipped their feature yet, they risked all bering in the "underperforming" category, simply to allow a peer team to have more "over performers" to reward them-- for actually delivering or for sufficiently sucking off the appropriate boss (or not getting knives in your back.)
There were a lot of (metaphorical) blowjobs and knives in the back in that organization.
Which is what you have to have when ALL of the management is non-technical.
You can't have a meritocracy when the people assigning ratings are incapable of judging merit.