A 360-degree review gathers observations from several perspectives. The employee, manager, peers and direct reports may see different situations. A report brings these views together; it is not an objective measure of a person’s worth.
Define the purpose first
Leadership development, collaboration feedback and formal performance assessment serve different purposes. Explain who receives the report and what decisions it informs. We recommend separating developmental feedback from automatic pay decisions. Do not promise a development-only process if the results will also inform employment decisions.
A meta-analysis of 24 longitudinal studies found generally small improvements in ratings after multisource feedback. A report cannot replace support, practice and follow-up.
Select people with relevant observations
Choose raters who have worked with the person during the stated period. Avoid selecting only friends or people with little opportunity to observe.
| Scope in our platform | Perspectives |
|---|---|
| 90° | Manager |
| 180° | Self and manager |
| 270° | Peers added |
| 360° | Direct reports added where applicable |
| 360°+ | Customers or partners added |
Providers use these labels differently. State the actual perspectives rather than relying on degrees. People without direct reports can still benefit from multisource feedback.
Design questions and pilot the process
Ask about one observable behaviour per item. Specify a reference period, such as the last three months. “Shares information needed to complete a task” is more specific than “is an excellent leader”.
Use a clearly anchored scale and an explicit way to record insufficient observation. In the current ready-made presets, mandatory rating items need a rating or an explicit N/A selection; comments are optional. A blank mandatory field is not an N/A answer. “Not enough opportunity to observe” is neither a midpoint nor a zero. Pilot the time burden, including how many colleagues each person rates. Around 20 questions can be a starting point, not a validity threshold.
See the 360 question template and competency framework guide. For multilingual teams, check equivalent meaning rather than translating labels mechanically.
Explain confidentiality before launch
State which perspectives are identifiable, who sees comments and when small groups are suppressed. Minimum group sizes reduce disclosure risk but do not guarantee anonymity. Comments and combinations of report filters can reveal identities. Read the confidentiality guide.
Interpret the report
Compare rater groups separately and consider their response counts. An average can summarise responses if its contents and weights are clear. A larger group should not silently dominate simply because it supplied more ratings.
A self–other gap invites discussion: were the situations, expectations or reference periods different? It does not automatically show a lack of self-awareness. A score alone cannot establish readiness for promotion.
Agree development actions
Discuss examples and agree a small set of priorities in an individual development plan. Repeat the review when there has been an opportunity to observe change. Everyday feedback need not wait for the next survey.
Create an account and prepare a review cycle. Review AI-generated question suggestions for relevance, clarity and purpose before use.
Department reports and the privacy of people being reviewed
A 360 review needs to protect both raters and the person being reviewed. A department chart is not automatically anonymous because names are absent: small teams and unique roles may identify the employee. Additional rater demographics can also expose feedback authors.
Do not count multiple ratings of one person as separate employees in an organisation report. Explain the unit of analysis, aggregation and rater-group weights. Comparing roles with different questions needs particular care. Group-result interpretation covers common risks; the privacy guide explains data boundaries and the ISO 10667 assessment-service standards.
Sources and scope
These sources support the principles discussed here. Examples and schedules are starting points to adapt, not standards or evidence that our templates have been validated.
