The annual review does not improve performance.

Almost every performance system sold to small businesses fails because it tries to do two unrelated jobs with one instrument.

Job one is performance improvement: getting more and better work out of the team. What delivers it is clear shared goals, a structure that fits the work, monthly structured debriefs, and pay that points the same way as the goals. It runs weekly and monthly. Structured debriefs alone are worth 20 to 25 percent.^A

Job two is legitimacy: is this workplace fair, clear and defensible? What delivers it is the performance review, done as a painless confirmation of feedback already given. It runs quarterly and annually. Every practitioner I have read agrees on this, and none of them can quantify it, so it gets a C.

Analysing an individual does not make a team produce more. What the review does, and what nothing else does, is make pay non-arbitrary, make expectations enforceable, give the employee due process, and create a record. Those are worth real money. They are simply a different product from improvement.

The failure mode is the annual review as a standalone event, done from memory. The manager rates on the last six weeks because that is what anyone remembers. The employee arrives unprepared because nobody recalls March in November. Both sides defend. Nothing improves, and the record is weak.

The fix is boring. A weekly log of three questions, 90 seconds, so the evidence exists before anyone is defending themselves. A prep sheet generated from twelve weeks of it, given to both sides a week before. Then the review is a confirmation. The money conversation is arithmetic, because the formula was published in advance. Nobody is surprised, and that is the whole point.