360 degree feedback: useful for development, dangerous for pay

360 degree feedback gathers views on someone from the range of people they work with — their manager, their peers, the people who report to them, sometimes customers — rather than from one line manager alone. The premise is straightforward and correct: a manager sees a fraction of how someone works, and the people alongside them see a different and often more accurate fraction.

It works well for development. It works badly, and sometimes destructively, when it is attached to pay or promotion, and the difference between those two uses is the whole subject.

Why 360 degree feedback surfaces things a manager cannot

• Peers see day-to-day collaboration: whether someone shares information, whether they are reliable under pressure.
• Direct reports see management behaviour, which is invisible upwards almost by definition.
• Cross-team colleagues see how someone behaves when there is no authority in the room.
• Several independent views separate a pattern from one person's impression.
• It catches the two failure modes single-manager review misses entirely: someone excellent upwards and difficult sideways, and someone quietly essential to everyone but their manager.

The predictable finding, across most organisations that run it, is a gap between how people rate themselves and how others rate them — and the size of that gap correlates with seniority. That is uncomfortable and it is the most valuable output of the exercise.

What happens when it affects pay

Once feedback influences a rating that influences money, the incentives invert. Reciprocal arrangements appear — you rate me well, I rate you well. Genuine criticism becomes risky, because a colleague may guess who wrote it and the professional cost is real. Managers gain a reason to steer who is asked. The data does not just get noisier; it becomes systematically biased in a direction nobody can measure.

The reasonable position is to keep 360 feedback developmental, shared with the person and their manager, and to leave performance and pay decisions to a process where the evidence is attributable and can be challenged. Organisations that ignore this usually rediscover it within two cycles, at the cost of the honest answers they had in the first round.

Running one that produces something usable

1. Say clearly, in writing, what it will be used for and what it will not.
2. Let the person suggest respondents, then have the manager add a few for balance — self-selection alone produces friendly panels.
3. Aim for six to ten respondents, enough that individual answers are not identifiable.
4. Ask about specific behaviours rather than traits: what should this person start, stop and continue.
5. Aggregate the numbers and pass written comments through lightly, removing identifying detail without sanitising the content.
6. Give the results in a conversation, never as a document sent by email.
7. Convert it into two or three actions, and revisit those before running the exercise again.
8. Do not run it more than once or twice a year — respondent fatigue is real and shows up as identical scores.

Anonymity, and its limits

Anonymity is what makes honest answers possible, and it is weaker than participants assume. In a team of five, a comment about a specific incident identifies its author regardless of the label on the form. Be honest about this rather than promising more than you can deliver: say that responses are aggregated and comments are passed on without names, and that in a small group people may still recognise a voice.

The related risk is the feedback used to settle scores. A moderation step — someone reading comments before they are passed on, removing personal attacks while keeping the substance — costs little and prevents the one bad experience that ends the practice.

Collecting it

Ettex Forms is the collection side: a form per person with the same behavioural questions, responses landing in one place, and a link sent to each respondent so nothing arrives as a reply-all thread. Responses can be reviewed together before anything is passed on.

It has no anonymity guarantee built for this purpose. Forms records who submitted where a respondent is identified, and it has no dedicated 360 module — no automatic aggregation into a report, no rater-group weighting, no separation of results from the person collecting them. For a small company that is usually fine with a trusted person handling it; for a formal programme where anonymity must be structural rather than promised, a dedicated tool is the right answer.

Keep it separate from the annual performance review even when the two run in the same month. The moment the feedback is read as an input to a rating, contributors start writing for the record rather than for the person, and the exercise stops producing anything useful.

Keep it developmental, ask about behaviours rather than traits, use six to ten respondents, deliver the results in a conversation — and keep it well away from the pay decision.