How to run a 360 feedback programme people actually trust

Most 360 programmes fail for the same reason, and it is not the questionnaire. It is that nobody was clear what the feedback was for.

Decide what it is for, and hold that line

A 360 used for development and a 360 used for appraisal are different instruments, even with identical questions. The moment raters believe their ratings affect someone's pay, promotion or job security, they rate differently — more generously, more cautiously, more politically. The data degrades, and it does not recover.

Decide before you start, say so plainly to everyone involved, and then behave consistently with it. If it is developmental, results belong to the focus person and do not go to their manager as a performance record. If you cannot commit to that, you are running an appraisal instrument and should design it as one.

The most common failure is a programme launched as developmental that quietly becomes an input to talent decisions. Raters notice, usually within one cycle, and the second round is worthless.

Anonymity is a design problem, not a promise

Telling raters their feedback is anonymous is easy. Making it true requires design.

Ratings should be reported by group — manager, peers, direct reports — rather than individually, with a minimum number of respondents per group before that group is reported at all. Three is a common minimum; below it, groups should be merged or suppressed.

Free-text comments are where anonymity usually breaks. People write recognisably, and a focus person can often identify an author from phrasing alone. Say so honestly in the rater instructions rather than promising a confidentiality you cannot deliver.

The manager's rating is a special case. There is usually only one, so it cannot be anonymous, and pretending otherwise fools nobody. Report it as the manager's view, openly.

Choosing raters

Focus people should nominate their own raters, with the list reviewed by someone else. Self-nomination alone invites a friendly panel; imposed lists produce raters with nothing useful to say.

Aim for people who have seen the person work recently and in enough depth to have a view. Six to ten raters across groups is usually sufficient. More raters do not produce better feedback; more relevant raters do.

Watch the response rate, and chase early. A programme that closes with half its raters unresponsive produces thin group averages, more suppressed groups, and a report the focus person can dismiss.

What happens after is the whole point

A 360 report handed over without a conversation is worse than no 360 at all. It creates anxiety, invites misreading, and generates no action.

Build the debrief into the programme rather than treating it as an optional extra. Someone trained should walk the focus person through the report, help separate signal from noise, and end with two or three concrete things to do. Small differences between scores are usually measurement noise, and an untrained reader will over-interpret them every time.

Then close the loop. Focus people who tell their raters what they heard and what they intend to change get better engagement next cycle. Raters who never learn whether their feedback mattered stop giving it thoughtfully.

Trust in a 360 programme is built by what happens after the report, not by the instrument.