The performance review that was decided before the form was opened

Everything here earned its place by changing an outcome. Nothing about annual appraisal ratings is included to round the number up.

The distribution usually exists before the ratings do

Pay and bonus budgets are set at an organisational level, which caps the total value of ratings that can be awarded across a population. That cap propagates downward as guidance about expected distribution, so managers arrive at reviews with a shape to fit rather than a blank sheet.

Calibration meetings then reconcile individual proposals against that shape, which is where a rating is genuinely decided. The one-to-one conversation happens either side of that process and is presented as its input, which is only partly true. This is not a conspiracy so much as ordinary budgeting, and it becomes corrosive only when the process is described as something it is not.

Forced ranking has a well-documented failure mode

Requiring a fixed proportion of low ratings assumes underperformance is evenly distributed, which is untrue in any team that has hired carefully. It also creates an incentive to avoid working with strong colleagues, since relative ranking makes a good peer group a personal liability. Several large organisations adopted and then abandoned strict ranking after concluding it damaged collaboration more than it improved performance.

Somewhere in the release notes, softer versions survive as guidance rather than as quotas, which reduces the worst effects and preserves most of the arithmetic. The evidence here is organisational and contested rather than experimental, so treat strong claims in either direction with some caution.

Recency bias eats the other eleven months

Memory for the most recent weeks is far stronger than for the same period a year earlier, and an annual review is written from memory. The result is that work completed shortly before the review counts more than identical work completed in the first quarter.

Written records dampen this, which is why keeping a running note of your own outcomes is the single most useful preparation available. It also explains the widely observed pattern of visible effort clustering in the weeks before review season. A manager who keeps contemporaneous notes produces measurably fairer reviews, and very few organisations make time for that.

Self-assessment is a genre with rules

The self-assessment is used as evidence in calibration, which means understatement is read as accurate rather than as modesty. Specific outcomes with numbers attached survive summarisation, while descriptions of effort and attitude do not. Framing contributions in terms the organisation already measures makes them legible, which is not the same as making them true or false.

People who find this distasteful are usually correct and are also disadvantaged by it, which is an uncomfortable pair of facts. Writing it once during the year, in a running document, converts an unpleasant annual task into a five-minute one.

Feedback and pay do not mix well

A conversation that determines pay is one in which the employee is defending a position, which is a poor state for absorbing development advice. Research on feedback generally supports separating evaluative and developmental conversations, ideally by weeks rather than by minutes. Organisations that split them report better engagement with the developmental part, which is the only part with any future value.

The bit they bury in the confirmation email: combining them saves manager time, which is the reason it persists despite being understood as a bad design. Where you cannot change the structure, mentally separating the two halves of the meeting is a reasonable individual adaptation.

One bad unit is bad luck; a pattern across thousands of buyers is a design decision.

Getting something useful out of it

Keep a running record of outcomes with dates and numbers, updated monthly, which takes minutes and eliminates the recency problem for you. Ask what would have been required for the next rating band, since a specific answer is actionable and a vague one is informative in another way. Ask when the rating was effectively decided, politely, because knowing the calendar tells you when the conversation should have happened.

Treat the developmental content as the valuable part regardless of the rating, as it is the only portion under your influence. And remember that a review measures how legible your work was to the organisation, which overlaps with its value without being the same thing.

Everything above, in order of what to do first

1. The distribution usually exists before the ratings do. Pay and bonus budgets are set at an organisational level, which caps the total value of ratings that can be awarded across a population.
2. Forced ranking has a well-documented failure mode. Requiring a fixed proportion of low ratings assumes underperformance is evenly distributed, which is untrue in any team that has hired carefully.
3. Recency bias eats the other eleven months. Memory for the most recent weeks is far stronger than for the same period a year earlier, and an annual review is written from memory.
4. Self-assessment is a genre with rules. The self-assessment is used as evidence in calibration, which means understatement is read as accurate rather than as modesty.
5. Feedback and pay do not mix well. A conversation that determines pay is one in which the employee is defending a position, which is a poor state for absorbing development advice.
6. Getting something useful out of it. Keep a running record of outcomes with dates and numbers, updated monthly, which takes minutes and eliminates the recency problem for you.

The takeaway

Keep the running record. It is the only part of the process you control.

It is not you being fussy. It is genuinely badly made.