A formal meeting to discuss and evaluate an employee's performance over a period.
A performance review is a structured, documented conversation in which a manager and employee assess work over a defined period against agreed objectives and set expectations for the period ahead. It is one event inside performance management rather than the whole of it: the review formalises judgements that should already have been made and communicated while the work was happening.
A performance review is a checkpoint inside a larger process. Performance management is the year-round activity: objectives set and revised, one-to-ones, feedback given close to the work, and a record of what was agreed. The review consolidates that evidence into a documented judgement at a fixed point. When the two are confused, the review is asked to do work it cannot do, and twelve months of silence followed by a single meeting is not a review but a verdict.
The working test is the no-surprises rule. If anything in the review is news to the employee, the failure happened months earlier, in feedback that was never given. A review that mostly confirms what has already been said in the ordinary course of managing is doing its job.
Organisations that abolish the formal review often find the judgement does not leave with it. Pay, promotion and progression still require someone to distinguish between people, and if the form no longer records that comparison it moves into a conversation the employee never sees, which is worse for both fairness and defensibility. The productive change is usually to make the event lighter and the year heavier, not to delete the event.
Most reviews resolve to a point on a scale, because points can be aggregated, compared and fed into pay models. The cost is that the number carries the rater's habits as much as the employee's performance. Variation between managers is frequently wider than the real difference in performance between their teams, so a rating is not comparable across an organisation until something has been done to make it so.
One distortion belongs to the form rather than the rater: the overall score produced by averaging dimensions. Someone rated highly on delivery and poorly on conduct averages to the middle, and the middle describes nobody. Composite scores hide the exact profile a manager most needs to act on, so keep dimension scores visible alongside any summary.
Forced distribution deserves a separate warning. Imposing a fixed curve assumes performance in every group is spread the same way, which is least likely to hold in the small teams where it is most often applied. A strong team must nominate someone as its weakest and a weak one is guaranteed its quota of top ratings. Both outcomes teach people that the rating describes the quota rather than the work.
| Distortion | What it looks like | What reduces it |
|---|---|---|
| Recency | The rating tracks the last few weeks of the period rather than the period | Contemporaneous notes kept through the year, not written up at review time |
| Central tendency | Almost everyone lands on the middle point and the scale stops discriminating | Behavioural descriptions for each scale point, plus a required example per rating |
| Leniency or severity | One manager rates a whole team high, another rates an equivalent team low | Cross-manager calibration before ratings are communicated |
| Halo and horns | One salient strength or failure colours every other dimension | Rate dimension by dimension across the team, rather than person by person |
| Forced distribution | A curve is imposed regardless of the real spread of performance | Use the distribution as a prompt to examine outliers, not as a quota |
A review is simultaneously a development conversation and an administrative input to pay, promotion and sometimes a capability process. These pull against each other. Honest discussion of weakness needs an employee willing to name their own gaps, and that same employee has an obvious reason not to when the conversation sets their pay. The usual mitigation is separation in time: the development review runs on its own cycle and the pay decision follows later on the evidence, so neither conversation deforms the other.
The administrative half has a longer tail than most managers expect. Where an employer later needs to show that a performance concern was real, raised, and given a genuine chance to be put right, the review record is the primary evidence. Years of unspecific "meets expectations" followed by a sudden problem is a weak account; specific, dated, contemporaneous notes are a strong one. Rating generously to avoid an awkward meeting is not a kindness, because it removes the employer's ability to act fairly later and denies the employee the chance to change something while there is still time.
Ask for the employee's self-assessment before the manager writes theirs, so the manager's draft does not become the anchor the employee merely reacts to. Calibrate across managers before ratings are communicated, never after: a retracted rating costs more trust than a delayed one. Define each scale point in observable behaviour rather than adjectives, because "exceeds expectations" otherwise means whatever the individual rater thinks it means. Close with a small enough number of objectives that they can actually be reviewed next time.
Treat the overall spread of ratings as data about the process, not only about the people. If one department returns no rating below the midpoint while a comparable department returns several, two rating cultures is a more likely explanation than two different populations, and the fix belongs to the process rather than to any individual.
Performance management is the continuous process: setting objectives, holding one-to-ones, giving feedback close to the work and planning development through the year. The performance review is one formal event within that process, where the accumulated evidence is consolidated into a documented assessment. A review with no performance management behind it is only a rating, and performance management with no review leaves no record.
Not always, but the judgement a rating stands for usually is. Pay, promotion and progression decisions require managers to distinguish between people, so if the form stops recording that distinction it is generally made elsewhere with no written trail. Dropping ratings works best when it is replaced by a documented alternative, such as written summaries and calibrated recommendations, rather than left as a gap.
There is no single correct cadence. Annual, half-yearly and quarterly cycles are all common, and the right interval depends on how quickly objectives and priorities change in the role. What matters more than frequency is that feedback is given close to the work in between, so the formal review confirms what the employee already knows instead of introducing it for the first time.
Explore our full platform with a 14-day free trial on Standard. Manage employees, post jobs, hire faster, and manage your tasks effortlessly with our all-in-one platform.