Systems & Organizations
Promote the Star, Kill the Engine
The better a salesperson's numbers, the likelier the promotion — and the worse their team sold afterward.
2026-08-312 min read
Laurence Peter's joke from 1969 — everyone rises to their level of incompetence — survived because it felt true, not because anyone had measured it. Alan Benson, Danielle Li and Kelly Shue measured it. Working from personnel records covering sales workers across a large panel of United States firms, they could see two things at once that are normally invisible: how well someone sold, and how the people under them sold after the promotion.
The two findings point in opposite directions. Sales performance strongly predicted who got promoted into management. Sales performance also predicted, negatively, how the promoted person's team performed afterward — the stronger the individual seller, the weaker the subsequent growth of the people they went on to manage. The firms were not promoting at random and getting unlucky. They were reading a signal that was real, legible, and precisely the wrong one, and the more of it a candidate carried, the likelier the mistake.
The interesting question is why a firm would keep doing this. A promotion is not only a staffing decision; it is the prize at the end of a tournament everyone else is still running. If the top seller is passed over, the message to the rest of the floor is that selling hard does not pay, and the effort of the many can be worth more than the competence of the one. Read that way, a company promoting its best individual contributor may not be making an error at all — it may be paying for effort with a management job, which is an expensive currency. That reframing changes the fix. If you are paying for effort, better promotion criteria will not help; what you need is a second currency — a technical ladder, pay that climbs without headcount, status that does not require direct reports — so the prize can be handed out without handing anyone a team.
Key insights
The cleanest signal you have about a candidate is the one least predictive of how they will manage.
Promotion doubles as a tournament prize, so until there is a second prize you will keep spending management jobs to pay for effort.
Why it matters
The evidence that makes a promotion feel obvious is individual output, which is the evidence least predictive of managing well — so your most confident promotions carry the most risk.
Without a second way to reward high performers, every promotion decision is quietly doing two jobs at once, and the staffing job is the one that loses.
Use this tomorrow
1List your last five promotions into people management and write beside each the single piece of evidence that most influenced it; count how many of those five were individual output.
2Take the two highest performers you are not planning to promote and write one sentence each on what they get instead this year; if you cannot finish either sentence, the technical ladder does not exist yet.
Go deeper
The study is "Promotions and the Peter Principle" by Alan Benson, Danielle Li and Kelly Shue, published in the Quarterly Journal of Economics in 2019. It draws on personnel data covering sales workers and their managers at United States firms — sales rather than software for a specific reason: individual sales output is unusually cleanly measured, and so is the output of the team someone inherits. That measurement clarity is the only reason the effect is visible at all. It is worth remembering when you hear the same result asserted for engineering; the finding is documented for sales, and the extension to other functions is an inference rather than a measurement.
Peter's own book, The Peter Principle, was written with Raymond Hull in 1969 and was satire. It proposed that hierarchies promote people until they stop being competent and then leave them there, which makes the whole thing sound like an accident of inattention. The measured version is less forgiving and more useful, because it suggests organizations are not failing to notice — they are optimizing for something else and accepting this as the price. A satire tells you to fix your promotions. A measurement tells you to decide what you are actually buying with them.
