Keith Rabois shares Peter Thiel's rule at PayPal: promote only when someone clearly outperforms their peers by a large margin. Promote too early and you risk demoralizing a high-potential person by putting them under someone else, hiding what they could actually do.
Chapters
So I actually learned from Peter thiel back at PayPal that you only really want to promote someone when there's a significant delta between that person's performance and the next best person. The bigger the delta, the easier it is to promote somebody. Otherwise, you're better off, as an executive leading an organization to leave things flat until the gap is just unmistakable. Then you can stitch things together. You don't really want to suppress someone's growth. So if you have a high potential executive leading an organization and you have 11 direct reports, artificially suppressing that person's potential by stitching them together under someone, layering them isn't really effective. You're going to demoralize them. You're never going to see how much upside they really have. And so as long as you think someone has unbounded upside, I'd be very leery of suppressing their potential.