The Case Against Annual Appraisals: What the Evidence Actually Says

Why this research matters

A clear-eyed review of the research on annual performance reviews - what they do well, where they consistently fail, and what the evidence supports as an alternative.

The $35 Billion Problem

CEB (now Gartner) estimated that the annual performance review costs large organisations an average of $35 million in management time annually for a 10,000-person company - and that in the majority of cases, this investment produced no measurable improvement in individual or organisational performance. More than a decade after this finding entered HR practitioner discourse, the annual review remains the dominant form of performance management in most organisations, despite a substantial and consistent body of evidence that it fails on almost every dimension it is designed to address.

This is not a case against performance management. The evidence in favour of clear goals, regular feedback, and structured development conversations is overwhelming. It is a case against one specific implementation - the once-a-year, high-stakes, backward-looking review - and an argument for understanding precisely which elements of the annual process create value and which destroy it.

What the Evidence Says Annual Reviews Do Poorly

The academic and practitioner research on annual performance review failures is remarkably consistent across different organisational contexts, cultural settings, and decades of study:

• Recency bias is structurally embedded: Managers evaluating a year's performance consistently over-weight the most recent one to three months. The result is that performance ratings reflect recent performance, not annual performance - systematically disadvantaging employees who performed exceptionally early in the year but experienced any difficulty late in it.

• Ratings are unreliable: Research published in the Journal of Applied Psychology consistently shows that 62% of the variance in performance ratings reflects the idiosyncratic standards of the individual rating manager, not the actual performance of the employee being rated. This is called the "idiosyncratic rater effect" and it makes cross-employee comparison of ratings largely meaningless.

• They activate threat, not growth: Social neuroscience research demonstrates that evaluation anxiety activates the brain's threat response, reducing working memory, narrowing cognitive perspective, and making people less likely to engage openly with feedback. The annual review - high-stakes, summative, and consequential - is structurally designed to produce exactly the psychological state least conducive to learning.

• The feedback gap is too wide: Research on feedback timing is unambiguous: feedback is most effective when delivered close to the behaviour it addresses. Annual reviews attempt to provide developmental feedback on behaviours that occurred eleven months ago - beyond the reach of effective behavioural change conversations.

What Should Replace Them

The evidence supports a different architecture: regular, brief check-ins (monthly or quarterly) focused on progress, obstacles, and near-term development; continuous, informal feedback normalised as part of everyday work; and a small number of structured conversations per year (two to three) focused on broader development goals and career direction. Pay and promotion decisions, where required, can be managed through calibration processes that aggregate manager assessments across the year - not through a single annual rating.

This is not a simpler system. It is a more demanding one - particularly for managers, who must develop the habit and skill of giving ongoing feedback rather than concentrating all evaluation activity into a single annual event. The transition requires significant manager development investment and a genuine shift in the organisational culture around feedback. But the evidence that it produces better outcomes - in engagement, development, and ultimately performance - is compelling and consistent.