Almost every coaching engagement opens with a 360. The best meta-analysis of multisource feedback finds performance gains so small they sit below the conventional threshold for “small” — and the wider feedback literature finds that more than a third of feedback interventions make performance worse.
Smither, London & Reilly, 2005
Personnel Psychology, 58
24 longitudinal studies
The question — does performance improve after multisource feedback — has a proper answer, from a meta-analysis of studies that followed people over time rather than asking them afterwards how it went.
| Whose ratings improved | Effect (d) |
|---|---|
| Direct reports | .15 |
| Supervisors | .15 |
| Peers | .05 |
| Self-ratings | −.04 |
For scale: the conventional benchmark for a small effect is .20. Every figure here falls below it, and self-ratings moved slightly backwards.
The authors state the practical implication without hedging: practitioners should not expect large, widespread performance improvement after employees receive multisource feedback.
38%
of feedback interventions
made performance worse
The landmark study here is not about 360s specifically — it is about feedback itself, and it is one of the most cited papers in organizational psychology. Kluger and DeNisi, in Psychological Bulletin in 1996, pooled 607 effect sizes from 131 studies, covering 12,652 participants and 23,663 observations.
The average effect was positive: d = .41, a moderate improvement. That number gets quoted constantly.
The next sentence does not. In their words: over 38% of the effects were negative.
More than a third of the time, giving someone feedback made their performance worse. And this was not an artefact of a few harsh studies — excluding the most extreme negative-feedback experiments still left roughly a third of effects negative.
Their explanation is that feedback which directs attention to the self rather than to the task tends to hurt. Tell someone their work has a specific fixable problem and they fix it. Tell them something about who they are as a leader and a good share of their attention moves off the work and onto defending themselves.
A 360-degree report is, structurally, a document about the self. That does not doom it — but it explains why it needs handling that most organizations do not give it.
The same meta-analysis identifies what separates the cases where behaviour improves from the cases where nothing happens. None of it is about which questionnaire you buy.
Purpose is the big one. Feedback used purely for development produced an effect of .28. Feedback feeding into administrative decisions — pay, promotion, performance ratings — produced .09. Roughly a threefold difference, from one design choice.
The recipient has to believe change is possible and has to accept that change is needed. Someone who disputes the premise gets nothing from the report, however accurate it is.
Something has to happen afterwards. Setting specific goals, talking the results through with someone, and taking visible action all predict improvement. A report delivered and filed predicts nothing.
If your organization runs 360s as part of the annual performance cycle, with results visible to the person who sets pay, the evidence says you are operating in the .09 condition and should expect close to nothing — while carrying the downside risk documented above.
Not because the tool is worthless. Because there is a better-evidenced way to answer the question a 360 is usually bought to answer.
A 360 asks people to rate the leader. The measures used here ask people about their own psychological resources and about who they actually go to for advice. Both are reports on the respondent's own state or behaviour rather than a judgement of someone else — which avoids the rating problems, and produces something the leader cannot argue with in the way they can argue with a score about themselves.
That design choice came out of the original study behind this practice, which measured servant leadership from the follower's side and asked what it was doing to them rather than what kind of leader they had.
Where a client already has 360 data, it gets used — carefully, developmentally, and never as the baseline. The conditions above are the ones that decide whether it helps, and they are contractible in advance.
Kluger, A. N., & DeNisi, A. (1996). The effects of feedback interventions on performance: A historical review, a meta-analysis, and a preliminary feedback intervention theory. Psychological Bulletin, 119(2), 254–284. Source of the 607 effect sizes, the d = .41 average, and the 38% negative finding.
Smither, J. W., London, M., & Reilly, R. R. (2005). Does performance improve following multisource feedback? A theoretical model, meta-analysis, and review of empirical findings. Personnel Psychology, 58, 33–66. Source of the per-rater effect sizes and the developmental-versus-administrative comparison.
Reviewed September 2026. Both papers were read in full text. If something on this page is wrong, write and say so — corrections are published.
The engagement the 360 is usually the opening move in.
Large effects · unstable constructThe other place a per-dimension profile gets handed to a manager on thin evidence.
Real construct · contested measurementAlso measured from the follower’s side — and it inherits some of the same problems.
Every construct this practice uses gets the same treatment, including the ones that make it look good and the ones that do not.