It is all relative though. If the team average performance is 50%, one can easily identify the very low and very high performers. The very low performers will b

It is all relative though. If the team average performance is 50%, one can easily identify the very low and very high performers. The very low performers will barely deliver anything at all, while the very high performers will deliver multiples of the average. Now - if the mean is higher, variance is low, and the distribution isn't symmetric, that's when things start to become harder. Worst case, you start firing people that are "low-performers" on paper, but in reality might be very close to the "meets expectation" workers. Which creates a very toxic environment, as your average workers will be walking on eggshells. And we know from history that this isn't some outlandish scenario in tech. There have been companies that have had(still have) a strict stacked ranking system - and come hell or high water, someone has to go.