"Well statistically half your employees by definition are sub-par." Not at all. First of all, 'par' doesn't mean the average of the skill levels of your employ

"Well statistically half your employees by definition are sub-par." Not at all. First of all, 'par' doesn't mean the average of the skill levels of your employees. 'Par' could be 'satisfactory' or 'the average skill level of the market as a whole'. Your own employees should not be a random sample of the market - otherwise why do you have a hiring process? Your employees should be above average - otherwise your hiring process needs improvement. And if you have a stand-out performer at either end of the scale, then that will skew the distribution. Thoughts: If you couldn't hire an A player in the first place, maybe you should think about replacing your hiring process or look to make improving people a strength? This improvement may even pay off on your hiring process in the future! It's difficult/impossible to judge the effect on a group dynamic that one person can have by any realistic rating system. Whatever system you choose will have exceptions and problems and you'll do some damage to your organisation by following it. "But if you can't hire replacements better than the people you fired, that's when you need to stop the practice." No, I disagree. Firing people and hiring them has cost. Your true indicator is 'if the risk of damage to your business caused by churning your employees and potentially not replacing them with much better candidates represents a larger lost profit than improving your staff (cheaper than hiring new), then you should stop trying to manage by silly quantitative rules and display the judgment and management for which you profess expertise.