360 degree feedback is a process where an individual is rated on behaviors and competencies by the people around them, typically their manager, peers, and direct reports, alongside a self-rating, and the results are compared. It is a development tool, not a performance rating. The research is unusually clear on that distinction, and unusually clear that using it for pay and promotion decisions is where it stops working.
How 360 degree feedback works
A 360 degree review collects ratings from several groups of people who see the same person in different contexts, then reports them side by side so the differences become visible. The standard rater groups are:
- Self. The person rates themselves on the same items everyone else rates them on. The gap between this and everyone else is often the most useful output of the whole exercise.
- Manager. The traditional downward view, and the only one a normal performance review captures.
- Peers. People at roughly the same level who see the day-to-day behavior a manager misses.
- Direct reports. The upward view, usually reported only in aggregate to protect anonymity.
Wider versions add team members at other levels, internal customers, and sometimes external parties such as clients or suppliers. Ratings cover behaviors and competencies (communication, collaboration, leadership, decision-making) rather than output, and results come back as a report showing each group's average against the self-rating. If you want to know what lands on your desk at the end of that, we walk through what a 360 degree feedback report contains section by section, including how to read the self versus rater gap.
360 degree feedback vs a performance review
These get confused constantly, and the confusion is what causes most 360 programs to fail. They measure different things for different purposes.
| 360 degree feedback | Performance review | |
|---|---|---|
| Who rates | Manager, peers, direct reports, self, sometimes customers | The supervisor, downward only |
| What is measured | Behaviors and competencies | Job requirements, objectives, results |
| Purpose | Development and self-awareness | Evaluation and administrative decisions |
| Typical output | A report the individual owns | A rating in the individual's file |
| Tied to pay or promotion | Should not be | Usually is |
What the research actually says
Most articles on this topic assert best practices without citing anything. The best available synthesis is Nowack and Mashihi's 2012 paper in the Consulting Psychology Journal, which reviews the evidence across fifteen practical questions. A few findings are worth knowing before you run a program.
Use it for development, not for decisions. Reviewing roughly two dozen longitudinal studies, Smither, London and Reilly (2005) found that improvement after 360 feedback was about three times larger when the process was used for development than when it was used administratively: an average effect size of .25 versus .08 across rater sources, excluding self-ratings. The effects overall are real but practically modest, so treat 360 as one input that compounds over years, not a lever you pull once.
Feedback does not automatically help. Kluger and DeNisi's 1996 meta-analysis, covering 607 effect sizes and 23,663 observations, found feedback improved performance on average (d = .41) but that performance actually declined in roughly one third of the studies. Feedback is an intervention with a failure rate, not a neutral gift.
Raters disagree more than you expect. Conway and Huffcutt (1997) found the average correlation between two supervisors rating the same person was only .50; between two peers, .37; between two direct reports, .30. Self-ratings correlate roughly .3 to .6 with other people's ratings and run significantly higher. Low agreement is normal and is not a sign the instrument is broken; it means people genuinely see different behavior in different contexts, which is the entire point of asking more than one of them.
People who overrate themselves take it worst. Brett and Atwater (2001) found negative feedback "was not seen as accurate or useful, and it did not result in enlightenment or awareness but rather in negative reactions such as anger and discouragement." Plan for the debrief, not just the report.
Comment volume matters. Smither and Walker (2004), following 176 managers over a year, found those receiving a small number of unfavorable comments improved significantly, while those receiving a large number relative to positive comments significantly declined. The study did not establish where the tipping point sits, so the practical rule is to curate comments rather than dump them.
How many raters do you need?
This is where vendor advice and research diverge sharply, and you should know which one you are following.
| Rater group | For statistical reliability (.70+) | Common vendor practice |
|---|---|---|
| Supervisors | At least 4 | 1 |
| Peers | At least 8 | 3 to 5 |
| Direct reports | At least 9 | 3 to 5 |
| Total raters | Rarely achievable | About 8 to 12 |
The reliability figures come from Greguras and Robie (1995), as reported by Nowack and Mashihi, who add the honest caveat that this standard "may not be practical" for a manager with three direct reports. Nobody assembles four supervisors. The useful takeaway is not to hit those numbers; it is to know that a typical 360 with one manager and four peers is a conversation starter with real measurement error, and to stop treating its decimal places as precise. Separately, 3D Group found that three or fewer respondents in a group may be inadequate for reliable measurement, which is the practical floor and the reason most tools refuse to report a group below that size.
One common worry is misplaced: Nieman-Gonder and colleagues (2006) found that raters chosen by the person being rated were as accurate as, or more accurate than, raters assigned to them. Letting people nominate their own raters does not automatically bias the result.
How many questions, and on what scale
Most 360 questionnaires run 20 to 40 items covering four or fewer competencies with three or four behaviors each. Keep it completable in 15 minutes: raters often fill in several of these in a cycle, and fatigue shows up as straight-lining.
On the scale, the evidence points somewhere specific. A 3D Group benchmark found five-point scales used 76% of the time and seven-point 16%, and Nowack and Mashihi conclude the "sweet spot" for 360 response scales sits between four and seven points. One design detail is worth changing if you have it wrong: Bracken and Rose (2011) found frequency scales ("never" to "always") are inferior to satisfaction or effectiveness scales, because frequency answers cluster and produce too little variation to tell anyone apart.
What is the difference between 180, 270, and 360 degree feedback?
The number refers to how far around the person you go. A 180 collects feedback from the manager and self, and sometimes peers, without the upward view. A 270 adds direct reports or peers but stops short of the full circle. A 360 covers manager, peers, direct reports, and self. The wider you go, the more context you capture and the more administrative weight you take on, which is why many organizations start at 180 for individual contributors and reserve the full 360 for people who manage others.
Why 360 degree feedback programs fail
The failure modes are predictable and mostly organizational rather than technical.
- It gets attached to pay or promotion. The moment ratings affect money, raters start managing the outcome rather than describing behavior. This is the single most-documented way to destroy the data, and it is why the developmental effect size is triple the administrative one.
- The report is delivered without a debrief. A PDF of uncomfortable numbers with no coach and no conversation produces defensiveness, which is precisely the state in which nobody changes anything.
- Anonymity is promised but not structured. If a manager has three direct reports and you report their group average, you have not protected anyone. Set a minimum group size, commonly 3 to 5, and say what it is up front.
- It gets gamed. Reciprocal rating arrangements are a well-documented problem at large companies running these at scale.
- Nothing follows. No development plan, no follow-up, no second round to see whether anything moved. The 360 becomes an annual ritual that costs goodwill.
There is also a real tension nobody resolves cleanly: follow-up conversations between the recipient and their raters have a strong effect on whether behavior changes, and strict anonymity makes those conversations impossible. Most organizations trade some anonymity for some usefulness. Just make the trade deliberately.
Frequently asked questions
What is 360 degree feedback used for?
It is used to build self-awareness and guide individual development, particularly for managers and leaders, by showing someone how their behavior is experienced by people above, beside, and below them. The evidence supports this use and does not support using it to set pay or make promotion decisions, where the measured improvement drops to roughly a third.
What are the four perspectives in 360 degree feedback?
Self, manager, peers, and direct reports. Some programs extend to internal customers or external clients and suppliers, but those four are the core. The value comes from comparing them: a person rated well by their manager and poorly by their direct reports is a specific, actionable finding that no single-source review would ever surface.
Is 360 degree feedback anonymous?
Usually, and it should be for everyone except the manager, whose rating is normally identifiable by definition. Anonymity is enforced by reporting only group averages above a minimum group size, commonly 3 to 5. Below that threshold the result is both identifiable and statistically unreliable, so most tools simply will not display it.
How often should 360 degree feedback be conducted?
Annually or every 18 months for most organizations. The constraint is not the survey, it is the gap needed for someone to actually work on something and for other people to notice. Running it more often produces rater fatigue and measures noise; running it less often means the feedback arrives disconnected from anything the person remembers doing.
How does 360 degree feedback differ from traditional performance reviews?
A performance review is a supervisor evaluating results against job requirements, and it typically feeds pay and promotion. A 360 collects behavioral ratings from multiple directions to help someone develop, and it should not feed those decisions. They answer different questions: the review asks whether you did the job, the 360 asks how it felt to work with you while you did it.
What competencies are measured in 360 degree feedback?
Typically communication, collaboration and teamwork, leadership, decision-making, accountability, and how someone develops the people around them. Good practice is to measure four or fewer competencies with three or four observable behaviors each, since a rater can describe what someone does far more reliably than they can assess an abstract trait. Where those competencies come from matters: a structured competency assessment defines the levels first, so the 360 rates people against a shared standard rather than each rater's private one.
Does 360 degree feedback actually improve performance?
Modestly, and only under the right conditions. Longitudinal research finds real but small improvements, roughly three times larger when the feedback is used developmentally rather than administratively, and a broader meta-analysis of feedback found performance declined in about a third of studies. It works when paired with a debrief, a development plan, and a follow-up round. It does not work as a report you email out.
Where 360 feedback stops and organizational assessment begins
360 feedback answers a question about one person. It cannot tell you whether the problem is the person at all. When four managers in a row score poorly on decision-making, the honest reading is usually not that you hired four bad managers; it is that decision rights are unclear, the process is undefined, or nobody has the information they need in time. That is an organizational finding, and no amount of individual feedback will surface it. Our companion piece on 360 degree feedback vs organizational assessment covers where the line sits and how the two fit together.
Assessmentcloud does the organizational half of that picture, not the individual half: it scores your whole company across culture and engagement, process maturity, digital maturity, skills gaps, and compliance readiness into one 0-100 report against the industry median, flat from $49 a month. To be plain about it, we do not run 360 reviews, and if individual development feedback is what you need, one of the dedicated 360 tools is the right purchase. If you want to know whether the pattern your 360s keep surfacing is actually a company problem, that is what an organizational health assessment is for. Results are diagnostic input, not a certified audit.
Sources
- Nowack, K. M., and Mashihi, S. (2012). Evidence-based answers to 15 questions about leveraging 360-degree feedback. Consulting Psychology Journal: Practice and Research, 64(3), 157-182.
- Smither, J. W., London, M., and Reilly, R. R. (2005). Does performance improve following multisource feedback? Personnel Psychology, 58(1), 33-66.
- Kluger, A. N., and DeNisi, A. (1996). The effects of feedback interventions on performance. Psychological Bulletin, 119(2), 254-284.
- Conway, J. M., and Huffcutt, A. I. (1997). Psychometric properties of multisource performance ratings. Human Performance, 10(4), 331-360.
RUN IT, NOT JUST READ IT
Score this dimension for your company
The interactive sample readout on the homepage shows exactly what you get: scored dimensions, industry benchmarks, and a prioritized action plan. Flat pricing from $49 a month.