When Everyone Passes, the Review Failed

The CEO emailed me his company's latest performance review distribution before our call. 47 employees. 43 rated meets expectations. 4 rated exceeds. Zero rated below. Zero. He asked, with apparent sincerity, whether it looked healthy. I told him it looked like a calibration exercise that had stopped calibrating.

I asked him to name three employees on the team who were genuinely underperforming. He named two within a minute. Both of them had received meets expectations. Their managers had filled out reviews that reflected what was politically possible to write rather than what was true. The review system was not collecting performance data. It was producing legal documentation that everyone had agreed to call performance data.

This is the most predictable failure mode of any review system that lacks forced distributions. Managers default to the higher rating because the higher rating is the easier conversation. Meets expectations requires no follow-up plan, no HR involvement, no awkward one-on-one. Below expectations requires a development plan, a conversation, a paper trail, and an emotional cost the manager would rather not pay. The math is consistent: the manager pays the cost of an accurate review once; the company pays the cost of an inflated review for the next eighteen months in the form of compounding underperformance that nobody can address because the paperwork says everything is fine.

The damage compounds at the top of the distribution. The strong performer who deserves exceeds notices, immediately, that meets and exceeds produce nearly identical compensation outcomes. They notice that the colleague who is, by any honest measure, underperforming received the same meets rating they did. They draw the correct conclusion: the rating does not measure performance. They calibrate their effort to the median, because the system is not rewarding the difference. Within two cycles, the strong performers have either left or quietly stopped pushing. The company has trained excellence out of itself.

Forced distributions are the fix that most companies refuse to implement because they are unpopular. The unpopularity is the feature. A managers' calibration session that produces ten percent below expectations ratings is uncomfortable, by design. The discomfort forces the conversation that the inflated review was avoiding, the kind of straight talk a team can trust. Most managers, faced with naming their lowest performer, will name an accurate one. The accuracy is what the system is supposed to produce.

The hard version of this discipline is not the bottom of the distribution. It is the top. Exceeds expectations has to mean something, which means it has to be rare, which means most managers cannot give it to most of their team. Managers will resist this. The strong performers across the company need the rating to mean something, which means a small percentage must receive it. The compensation that follows must visibly differ from the median.

The review system you have is the performance distribution you will produce. If your reviews compress everyone toward meets, your team will compress its performance toward the median over time. If your reviews produce a real distribution, the team will too. The choice between comfort and calibration is, in practice, the choice between mediocrity and the company you wanted to build.