Employee calibration is broken

Most big companies have an employee calibration process to make sure managers are on the same page about career levels, performance ratings, and promotion readiness.

Many of those companies do calibration synchronously. A bunch of managers get on a big scary meeting and present their direct reports’ performance reviews to each other. Then they all debate about whether Fred’s measurable results earned his rating and whether Susan deserves the promotion she’s up for.

In theory, calibration ensures that everyone ends up with the rating they deserve based on the work they did. But in practice, calibration tests the managers more than the employees. Two identical employees could have wildly different ratings based on their manager’s presentation skill, willingness to argue, political sway, or opinions about what the ratings mean.

Some managers are just more talented than others at answering the tough on the fly questions that come up in calibration. Some managers can be seen “whipping votes” for their direct reports’ promotions in the weeks before calibration. Some managers have been around long enough to point at precedents set in past review cycles. Some managers are freaking intimidated by being in a room with 20+ peers and superiors.

Plus, psychological safety in calibration tends to be low low low, for a couple reasons:

1. Everyone thinks their own direct reports deserve high ratings and promotions, but that’s usually a zero sum game, especially if there’s an enforced curve. So everyone is fighting for the same thing while pretending that they are working together to give people the ratings they deserve.
2. The managers in calibration are often not a jelled “first team” with each other. In some cases, they’re in calibration with people they haven’t talked to since the last calibration. There’s no mutual trust and respect built up to lean on.

None of these things are fair to the employees being evaluated.

These are all hard problems, and I have no silver bullet. But I do know one thing: making calibration asynchronous helps a lot. There should be no “presentation” in calibration. All managers should be able to read the performance reviews and comment on them. Then there should be plenty of time for asynchronous discussion, and maybe dot voting.

No more rapid fire Q&A with zero time to think. No more sweaty palms from thoughts like “OMG ten directors are staring at me.” No more ratings earned by being charismatic or argumentative as opposed to the work itself.

Calibration will always be broken, but making it asynchronous fixes some parts of it.