360 degree feedback: how it works, what to ask, and when it changes behaviour

A regional sales director in Riyadh opens her first 360 report. Her team rated her a full point lower than she rated herself on "listens before deciding". Two weeks earlier, HR had told every manager that the scores would feed the year-end bonus. In the next cycle, almost every score in the company went up.

The scores improved. The behaviour did not.

The short answer: 360 degree feedback is a process for collecting, quantifying and reporting coworkers' observations about one person: their manager, peers, direct reports, sometimes clients, plus a self-rating. A cycle runs in steps: agree the purpose, choose the behaviours, pick raters, collect ratings confidentially, report by rater group, then set goals and act. On average the improvement that follows is small. It is larger when the feedback is used for development only, and when people set goals and discuss the results with their raters.

What 360 degree feedback is, and how it works

Bracken, Rose and Church (2016) define 360 feedback as a process for collecting, quantifying and reporting coworker observations about an individual. In their definition it captures how often people see specific behaviours, allows meaningful comparisons across raters and over time, and leads to lasting change in behaviours the organisation values. Each rater group sees a different part of the job.

The "three or more" rule comes from practice: Bracken and colleagues note that ratings from direct reports, peers and clients are almost always protected this way.

How does 360 feedback work? The steps of one cycle

1. Agree the purpose in writing. Development only, or also decisions. 2. Choose 12-15 behaviours that matter for the role. 3. Pick raters. The person proposes, their manager approves. 4. Brief raters on what each behaviour means and how the scale works. 5. Collect ratings through a confidential system. 6. Report by rater group, never below the minimum group size. 7. Debrief with a trained coach or HR partner. 8. Set two or three goals, no more. 9. Share the goals with raters and ask for suggestions. 10. Repeat after 6-12 months and check the goal behaviours.

What the research says 360 feedback can and cannot do

The honest summary: it works sometimes, modestly, and under conditions you control.

• The average gain is small. Smither, London and Reilly (2005) pooled 24 longitudinal studies. Corrected mean improvement was 0.15 standard deviations in direct report ratings, 0.05 in peer ratings and 0.15 in supervisor ratings. By the usual rule of thumb, 0.20 is small.
• Feedback can backfire. Kluger and DeNisi (1996) analysed 607 effect sizes from 23,663 observations of feedback in general. Feedback improved performance on average, but over a third of the interventions decreased it. Effects weakened when feedback drew attention to the self rather than the task.
• More sources do not add much. The CIPD's 2022 evidence review found that gains over time are small, and that feedback from more sources made no difference compared with feedback from direct reports only.
• The question is still open. A 2026 protocol for a new Campbell systematic review notes that researchers still disagree on whether multisource feedback improves performance.

What predicts improvement is what people do next. Smither and colleagues found it most likely when recipients see a need to change, believe change is feasible, set goals and take action. In a five-year study of upward feedback, managers who met their direct reports to discuss the results improved more, and improved more in the years they held that discussion. Change can take more than one round: in that study, ratings did not move between the first and second year, but rose after.

Development use vs appraisal and pay

Smither and colleagues found larger effects when feedback was used only for development, with a caution: very few studies used it for administrative purposes. Their review also cites evidence that ratings collected for administrative purposes were more affected by rater bias, and that when ratings counted and people could choose their raters, they chose raters they liked.

The CIPD review makes the same point about appraisal in general. Managers rate people very differently for pay than for development, and using the same ratings for both is likely to be unfair. Its advice: separate the two conversations.

There is a counterpoint. Bracken and colleagues argue that "development only" is a misnomer, because even development programmes end up informing talent reviews. They cite benchmark data showing companies using 360 for performance management rising from 27% in 2003 to 48% in 2013 (3D Group). Their answer is to set the purpose at design, and they warn that changing it midway carries ethical, cultural and legal risks.

The obvious answer, and why to test it

The obvious answer is to tie 360 scores to raises, so people take them seriously. Test that before you announce it.

The scores will move, but not always for the right reason. When ratings carry money, raters have reasons to inflate them and ratees to choose friendly raters. A higher average may mean better behaviour, or better politics.

Participation can drop before you start. The clearest published regional case we found is a 2016 study of 30 emergency physicians at a military teaching hospital in Bahrain. When multisource feedback was introduced, physicians declined to sign consent or take part, for fear it would be used against them in promotions and contract renewal. Being rated by nurses was also not widely accepted. The hospital went ahead after briefings and an official letter from the Chief of Medical Staff stating that the feedback would be used solely for improvement. This is one hospital and one profession, so treat it as a warning, not a forecast.

A fairer test. Run one development-only cycle. Then let 360 inform decisions in one narrow way: did the person set goals, share them and act? That rewards using the feedback, not the score.

Self-other gaps: why your rating and theirs disagree

Expect the self-rating to differ. That is normal, and it is the most useful part of the report.

• Agreement is modest. In Heidemeier and Moser's (2009) meta-analysis, self and supervisor ratings correlated at 0.22 (0.34 after correction), across 115 samples and 37,752 people. In Western samples, self-ratings were more lenient.
• Culture changes the gap. Atwater and colleagues (2009) studied 964 managers from 21 countries. Self and other ratings agreed more in countries high on assertiveness and on power distance. Ratings were more lenient in individualistic cultures. None of the 21 countries was in the Arab region, and we found no comparable study for Egypt, Saudi Arabia or the UAE. Do not assume how your managers rate themselves; look at your own data.
• The gap can drive change. Smither and colleagues cite two studies in which people who overrated themselves at first improved more than others afterwards.

Read the gap by rater group. A manager whose boss and peers agree with her, but whose team does not, knows which conversation to hold.

360 feedback questions: a ready set of 15

Rate how often the person does each one: 1 = almost never, 5 = almost always, plus "not able to judge".

Communication 1. Explains the reasons behind decisions that affect the team. 2. Listens to a different view fully before responding. 3. Shares information people need before they need it.

Decisions and delivery 4. Sets clear priorities when everything seems urgent. 5. Does what they said they would do, by the date they agreed. 6. Makes decisions in time for others to act on them.

Developing people 7. Gives specific feedback within a week of the work. 8. Gives team members work that stretches their skills. 9. Recognises good work by naming what was done well.

Collaboration 10. Involves other teams before decisions that affect them. 11. Raises disagreement directly with the person concerned. 12. Shares credit with the people who did the work.

Fairness and trust 13. Applies the same rules to everyone in the team. 14. Admits mistakes openly. 15. Responds to bad news without blaming the messenger.

Add two open questions: "What should this person keep doing?" and "What is one thing they should start or change?" They give you strengths and weaknesses in the raters' own words. Ask for behaviour, not judgment: "talks over others in client meetings" is usable; "arrogant" is not. Items 11, 14 and 15 overlap with psychological safety at work. For other ready-made item sets, see the employee survey templates.

How to keep it anonymous enough in small teams and hierarchical cultures

Most 360 tools are confidential, not truly anonymous: the system knows who rated but reports only group averages. Our guide to whether employee surveys are anonymous explains the difference. In small teams, that promise is easy to break by accident.

• Set a minimum group size, and keep it. Three raters per group is the common floor. If a manager has two direct reports, do not show a direct report score.
• Merge groups rather than expose people. Combine peers and direct reports into one "others" score when either is too small.
• Tell people the manager's rating is visible. It is usually reported alone, so it is not anonymous. Say so before anyone rates.
• Edit comments for identifiers, not meaning. Remove names and details only one person would know.
• Let a neutral party assign raters. The Bahrain hospital's independent team picked them at random from people who had worked with each physician for at least two months.
• Put the purpose in writing, from someone senior, as the Bahrain hospital did.
• If upward rating is still too risky, ask the team as a group. A facilitated focus group without the manager in the room can collect the same themes.

Your next step

This week, write one sentence stating who will see each 360 report and what it will and will not be used for. Have your leadership team sign it before any rater is invited. Then count the raters each person would have per group, and merge any group below three.

Frequently asked questions

What does 360 degree feedback mean?

It is feedback on one person's behaviour from the people around them: manager, peers, direct reports, sometimes clients, plus a self-rating. Results are reported by rater group, so the person can compare how different people see them. It measures observed behaviour, not personality.

How does 360 feedback work?

Agree the purpose, choose 12-15 behaviours, pick and brief raters, collect ratings confidentially, report by rater group above a minimum size, debrief, and set two or three goals. Studies find more improvement when people discuss results with raters and act on them. Repeat after 6-12 months.

What questions should a 360 feedback template include?

Group 12-15 behaviour-based items by competence, such as communication, decisions and delivery, developing people, collaboration, and fairness. Rate how often each behaviour happens, with a "not able to judge" option. Add two open questions: what to keep doing and what to start or change.

What are good 360 feedback examples of strengths and weaknesses?

Good examples name a behaviour and a situation. A strength: "explains the reasons behind deadline changes, so the team can plan." A weakness: "decides in meetings before hearing the people who will do the work." Avoid labels like "great leader" or "difficult", which give the person nothing to act on.

Should 360 feedback be anonymous, and should it decide pay?

Peer and direct report ratings should be confidential and reported only for groups of three or more; the manager's rating is usually visible, and people should know that. On pay, the research is cautious: in one meta-analysis, gains were larger when 360 was used only for development. Decide the purpose before launch and keep it.

Sources

• Bracken, D. W., Rose, D. S. and Church, A. H. (2016), "The evolution and devolution of 360° feedback", Industrial and Organizational Psychology 9(4), 761-794. doi.org/10.1017/iop.2016.93
• Smither, J. W., London, M. and Reilly, R. R. (2005), "Does performance improve following multisource feedback? A theoretical model, meta-analysis, and review of empirical findings", Personnel Psychology 58(1), 33-66. doi.org/10.1111/j.1744-6570.2005.514_1.x
• Kluger, A. N. and 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. doi.org/10.1037/0033-2909.119.2.254
• Heidemeier, H. and Moser, K. (2009), "Self-other agreement in job performance ratings: a meta-analytic test of a process model", Journal of Applied Psychology 94(2), 353-370. doi.org/10.1037/0021-9010.94.2.353
• Atwater, L., Wang, M., Smither, J. W. and Fleenor, J. W. (2009), "Are cultural characteristics associated with the relationship between self and others' ratings of leadership?", Journal of Applied Psychology 94(4), 876-886. doi.org/10.1037/a0014561
• Cioca, I. and Gifford, J. (2022), "Performance feedback: an evidence review. Practice summary and recommendations", Chartered Institute of Personnel and Development. cipd.org
• Wietrak, E., Cioca, I., Gifford, J. and Urbano, C. (2026), "PROTOCOL: Multisource feedback and work performance: a systematic review and meta-analysis", Campbell Systematic Reviews 22(2). merit.url.edu
• Al Ansari, A., Al Meer, A., Althawadi, M., Henari, D. and Al Khalifa, K. (2016), "Cross-cultural challenges in assessing medical professionalism among emergency physicians in a Middle Eastern country (Bahrain): feasibility and psychometric properties of multisource feedback", International Journal of Emergency Medicine 9, 2. doi.org/10.1186/s12245-016-0099-2