That's also the case at Amazon. And likely most large organizations. It's not a strict stack rank but an expectation that over a certain population size, there is a certain percentage of under-performers. It's also used as a method of calibration. Performance evaluations are hard.