Well, given that performance reviews are usually subjective, it’s important to calibrate those within a business unit to reduce this discrepancy. If A tends to

Well, given that performance reviews are usually subjective, it’s important to calibrate those within a business unit to reduce this discrepancy. If A tends to be a tough grader and B tends to be an easy grader things must be normalized to provide equal rewards for equal performance. Otherwise B’s people get all the money. This also serves to train in the new managers into the process and approaches to evaluation. However, if it’s the kind of stack ranking that must yield one “fire” and two “must improve” then it’s evil to use. Imagine, somehow, that a group has all legitimate superstars. One gets tossed and at least two leave with the first review cycle. The rest, being observant, probably leave over the next 6-12 months after such a surprise. This turnover would be perfectly reasonable for a website but somewhat fatal for, say, an auto-driving group.