2026 Guide

How to design a 360 review

The four voices and what each one sees, why anonymity does not apply equally to all of them, the wording detail that ruins most questionnaires, and how to collect feedback from your own clients.

The short answer

A 360 does not measure performance. It measures the gap between how a person sees themselves and how others see them. The average of all responses is the least useful number in the entire report: what opens the coaching conversation is the difference.

And above all, two asymmetries. What the person scores high and their environment scores low, which is usually a blind spot. And what they score low and their environment scores high, which is nearly always a strength they do not recognise in themselves.

The four voices, and what each one sees

None of the four sees the whole picture. That is why you need all four, and why they cannot be blended into a single average.

Always published
The person
The self-assessment. It does not add another truth: it provides the reference point everything else is compared against.
Why it is always published: it is not a group, it is the person looking at themselves. There is nothing to protect, because they already know what they answered.
Published and flagged
Their manager
Sees results, decisions and behaviour under pressure. Does not see the day to day, nor how they treat people they do not need to impress.
The rule: always published, but the row is flagged as a single voice when only one person answered. A 360 without the manager's view is half a tool; what you cannot do is publish it without saying how many people it came from.
Threshold applies
Their peers
How they collaborate, negotiate and behave when nobody holds authority over anybody. The voice that stings most and gives most.
The rule: not published below a minimum number of responses. With three, working out who said what is a matter of looking twice.
Threshold applies
Their team
The only group that sees how the person behaves when they hold power. Nobody else can answer that.
The rule: the same minimum as peers, and with more reason: here the respondent has something to lose if identified.

The underlying mistake is applying the same threshold to all four groups. It sounds fair and causes two harms: either the manager's voice gets suppressed, since it is nearly always one person, leaving the 360 half done — or it gets published without warning that it came from one person, and the person assessed reads their manager's view as «the organisation».

The detail that ruins most questionnaires

Every question has to be written twice: once for the self-assessment and once for whoever is answering about another person. Reusing the same sentence is the commonest failure when building a 360.

For themselves
«I delegate my team's work well.»
For everyone else
«They delegate their team's work well.»

They look like the same question and they are not. The second requires the respondent to have seen delegation happen; the first only asks for an opinion about oneself. If a peer has never been in that situation, a single wording forces them to invent an answer, and that noise ends up inside the average.

Practical rule: if a question cannot be written in the third person without sounding odd, it is not observable. And what is not observable should not be in a 360.

The five steps to building one

Say what it is for, before writing anything
A developmental 360 and a 360 feeding a promotion decision are not answered the same way. If it is not said up front, people assume the second and answer defensively. Everything else depends on this step.
Choose the voices and how many per group
Fewer than three peers leaves the group below the threshold and unpublishable. More than eight and nobody answers carefully. Between four and six per group is usually where people still take it seriously.
Write every question twice
One wording for the self, one about the person. And discard every question that is not observable from outside.
Set the threshold per group, not globally
Self-assessment always visible. Manager visible but flagged. Peers and team with a minimum. And decided before the questionnaire goes out, not once the answers are in.
Give it back by looking at the gap
The debrief session does not start with the average. It starts with the two questions showing the largest distance between the self-assessment and everyone else, in both directions. That is where the work is.

The other feedback: your own client's

A 360 assesses the person being coached. This is the opposite: what your client thinks of your work. It is collected far more simply, and almost nobody does it systematically.

What CoachPro Tools has, and what it does not

The 360 exists, inside the Organisations module. Several people answer about one person, and the result is grouped by the person assessed and by rater type — not by department, which is a different grain entirely. Each instrument knows whether it is answered about oneself or about someone else, which is why it asks for the second wording of every question.

The anonymity rule is the one described in this guide, because it is the one implemented: self-assessment always visible, manager visible and flagged when it is a single voice, peers and team with a configurable threshold. And 360 responses stay out of the departmental average: mixing a leader's self-assessment with their peers' view produces a figure that means nothing.

What is not there, and worth knowing before signing up: bringing third parties into an ordinary one-to-one process. If you work with an individual client and want their manager's or peers' view without setting up an organisation module, that function is still pending. The 360 lives in Organisations, not on a standalone client record.

For feedback work inside the session there are standalone instruments: individual 360 feedback, the SBI model, the feedback burger, the Johari window and group performance feedback. Plus the process satisfaction survey, with its two permissions.

Frequently asked questions

How many people does a 360 need?

Between four and six per group. Fewer than three peers puts the group below the threshold and it cannot be published. More than eight and people start answering on autopilot.

How often should it be repeated?

Once a year, or at the start and end of a coaching process of six months or more. Repeating it quarterly exhausts the respondents and leaves no time for anything to change.

Should the coach see individual responses?

Not if you promised anonymity. The coach sees what the person assessed sees, grouped. Seeing who said what turns the debrief into an investigation, and next round nobody answers honestly.

What if a very harsh comment comes up?

You give it back, but not first and not alone. An isolated comment is not a pattern, and presenting it as one makes the person defend themselves against that sentence instead of looking at the whole.

Is a 360 useful for someone who works alone?

Not very. With no peers and no team you are left with two voices, and with two there is no 360: there is a comparison. For an independent professional, a survey to their own clients works better.

Keep reading

The gap, not the average

360 grouped by person assessed and by rater type, with an anonymity threshold per group. Inside the Organisations module. From 30 USD a month.

See plans from 30 USD/month → Or see how the Organisations module works

30-day guarantee · No contracts · Instant access · One-off payment available