How does a 360° leadership assessment work? Six phases, a realistic timeline and three ways it goes wrong

The six phases at a glance

A 360° assessment is a process, not a questionnaire – and most of what goes wrong happens before the survey opens or after the report lands, not in between. If you want the concept first, start with What is 360° feedback?. This page is about the sequence: which phase demands which decision, and how long each realistically takes.

Definition360° feedback as a process

Bracken, Rose and Church (2016) define 360° feedback not by the instrument but by what the process must enable: collecting rater perceptions of specific behaviours, allowing meaningful comparisons between rater groups and over time – and sustainable behaviour change that the organisation values. A process that ends at the report fails the third test.

Phase by phase: the decisions that matter

1. Preparation: purpose before tooling

The first decision is the one that matters most: is this programme for development, or does it feed appraisal, promotion or pay? Both are possible, but not at once – and the answer shapes how candidly raters answer and how openly the leader listens. Preparation also fixes the instrument and the rater groups (supervisors, peers, direct reports), settles the anonymity rules and prepares the communication. Data protection belongs in this phase, not at the end; the questions to settle are laid out in GDPR and 360° feedback.

2. Nomination: who gives feedback, and how many

The research on rater numbers points one way: reliable group scores need more raters than most programmes invite (Greguras & Robie, 1998). At the same time, a leader can only nominate people who actually observe them at work. The compromise is to set minimums per group, choose for proximity of observation rather than sympathy, and cap the number of assessments any one rater is asked to complete, so the same colleagues do not end up in five processes at once. Practical criteria are in How to select raters; the numbers question in How many raters does a 360° need?.

3. Feedback window: short, clear, with reminders

Two to three weeks works well in practice: long enough to absorb holidays and project peaks, short enough to keep the topic alive. A questionnaire that takes 15 to 20 minutes gets completed; one that takes 40 gets postponed. Reminders should be automated and measured, and the decision on extensions should be taken before the window opens – an ad-hoc extension teaches everyone that deadlines are negotiable.

4. Aggregation and review: anonymity is not a formality

This is where the promise made in phase 1 is kept or broken. Group scores should appear only when enough responses exist; free-text comments must not be attributable to a person. Whether the system enforces those rules or someone applies them by hand is a difference worth clarifying before you buy – see Is 360° feedback anonymous?. A review by an experienced consultant before release also catches reports that would be misread without context. What a good report should contain is covered in What a 360° report should contain.

5. Debrief: the conversation is the product

The report is material; the conversation is the value. A good debrief does not start with the lowest number but with the pattern: where do self- and other-ratings agree, where do they diverge, and what might sit behind that? Who leads the conversation – an external consultant, an internal coach or HR – shapes how openly the leader deals with surprises. The trade-offs are set out in Consultant-led or self-service?.

6. Follow-up: turning information into feedback

Without follow-up, a 360° report is information, not feedback. Two or three concrete development goals, one person who asks how it is going, and a date for the second round are the minimum. A repeat shows whether observed behaviour has actually changed – and turns the programme from a one-off measurement into a process with a sequel.

Three things that derail 360° programmes

Rater fatigue

Rater fatigue sets in when the same colleagues are asked in several processes at once, when the questionnaire is too long, or when rounds overlap. The result is low completion and fast, undifferentiated answers – both of which undermine exactly the reliability you needed more raters for. The remedies: a cap per rater, a compact questionnaire, bundled invitations, and one window per cohort.

Unclear purpose

Feedback is not automatically helpful. Kluger and DeNisi’s (1996) meta-analysis of 607 effect sizes found that feedback interventions improved performance on average (d = 0.41) – but more than a third of them made performance worse, which the authors’ theory attributes above all to attention shifting from the task to the self. A programme without a clear development purpose produces exactly that shift: raters hedge, leaders defend, and the report becomes a verdict instead of a starting point.

No follow-up

Smither, London and Reilly (2005) analysed 24 longitudinal studies of multisource feedback: average improvement in others’ ratings over time was small, and the authors explicitly warn practitioners not to expect large, widespread gains. Improvement was more likely when recipients recognised a need to change, believed change was feasible, set goals and acted – working with a coach, discussing the feedback, doing development work. In short: the report alone moves little; what happens afterwards decides.

How LEADBeyond 360° runs the process – one example

1. Preparation: objectives, cohort, rater groups, anonymity threshold, timeline and language (German or English) are agreed with the consultant. HR can create a cycle and propose process settings; methodology settings stay with the consulting team.
2. Nomination: the leader, the consulting team or both nominate from a prepared pool; recommended minimums (default: two supervisors, two peers, two direct reports) are configurable per cycle. Invitations go out bundled at the end of the nomination phase – each rater receives one invitation listing every leader they are asked to rate.
3. Feedback window: self-assessment in about 18 minutes, rater feedback in 10 to 15 minutes, with autosave and the option to decline an assignment. Reminders run automatically after 3, 7 and 14 days plus a warning 48 hours before the deadline – configurable and can be switched off.
4. Aggregation and review: group scores appear only once at least the anonymity threshold (default: two) is reached; below that they are folded into "all others". A report is generated once the self-assessment is complete, at least two rater groups meet the threshold, and the deadline has passed or every invited rater has submitted. A consultant reviews every report, can add annotations, and releases it – manually or on a pre-set go-live date.
5. Debrief: the report with twelve interactive views and PDF export is the basis for the development conversation. The stress signature and predicted blind spots are marked in it as hypotheses for the conversation, not as diagnoses.
6. Follow-up: released reports stay unchanged; late responses create a new version that only becomes visible after an explicit re-release. A second round runs as its own cycle, and the report shows the change against the first. Throughout, HR sees progress and completion rates, never individual results.

What the model behind it measures – Belief, Capability, Trigger, Behavior, Outcome and Well-being – is described on the methodology page.