What is multi-rater feedback?
Multi-rater feedback gathers perspectives on a person’s work from several people who have relevant experience of it. These may include a manager, peers, direct reports, customers and the person’s own self-assessment. A 360-degree process commonly uses several of these directions.
The purpose is to build a useful development picture: strengths to continue, behaviors to improve and differences in how the work is experienced. Agreement can be valuable, as can disagreement. The exercise should not manufacture a gap simply because several groups were invited.
Ratings and comments contribute different evidence. Well-designed ratings can reveal patterns across defined behaviors; comments can supply examples and context. Read them together instead of treating either as automatically more truthful or useful.
Define the purpose before collecting feedback
Explain why the exercise is being run, who will receive the results and how they will be used. A developmental conversation and a formal employment decision are different purposes. Do not invite candid developmental feedback under one arrangement and later use it for an undisclosed purpose.
Choose the relevant behaviors and the time period people should consider. A broad item such as “is a good leader” gives raters room to interpret different things. A clearer behavior, such as making priorities explicit before work begins, gives the feedback a more practical focus.
Plan a supported review and follow-up, not only the report delivery. CCL’s guidance emphasizes clear outcomes, trust and explicit confidentiality arrangements when implementing a 360 initiative. See its implementation guidance.
Choose raters with relevant experience
Select people who have had a reasonable opportunity to observe the behaviors being assessed. A long list of names is not a substitute for relevant exposure. A colleague who rarely works with the person may not be able to assess day-to-day delegation or support.
Document how raters are selected and avoid a process that only invites supporters or only seeks criticism. Include an appropriate way to say a behavior was not observed. Treat that response as missing evidence for the item, not a low rating.
The appropriate groups depend on the role. A project lead without direct reports may need collaborator perspectives rather than an artificial direct-report category. A manager may receive different evidence from peers and reports because those groups observe different situations.
Ask about behavior and examples
Scroll horizontally to see all columns →
| Focus | Rating item | Open follow-up |
|---|---|---|
| Clarity | Makes priorities clear before work begins. | Describe a situation where priorities were clear or unclear. |
| Listening | Seeks relevant views before making a decision. | What did you observe that influenced your answer? |
| Follow-through | Communicates when an agreed commitment changes. | What should this person continue or change? |
These are illustrative items, not a validated assessment instrument. If you need a validated measure or formal norms, select and use an appropriate instrument under its conditions. Do not imply that a custom questionnaire has the same measurement properties.
Define the scale and reference period. Avoid combining several behaviors in one rating. Ask for useful examples without inviting unnecessary personal information or speculation about motives.
Read the report in a useful order
- Review the purpose, scope, response coverage and reporting rules.
- Look for strengths and development needs that recur across evidence.
- Compare perspectives where reporting is appropriate.
- Read comments and ratings together, including contradictory examples.
- Select a small number of practical development priorities.
An overall average can summarize some information, but it should not conceal relevant group patterns. Group means also need context: how many valid responses contributed, what raters could observe and whether variation exists within the group.
A chart is not inherently a poor report. The important question is whether the participant understands the evidence and can use it. A text-only summary can be equally unhelpful if it hides uncertainty or substitutes generic advice for what people actually said.
Worked example: interpret a perspective gap
Consider a fictional project lead, Alex. Feedback from a manager describes timely decisions. Peer comments describe clear project direction. Several direct-report comments describe limited opportunity to question a plan before it is finalized.
Those accounts can all be true. Different groups may experience different parts of the work. The useful development question is whether Alex can preserve decisiveness while making space for relevant input earlier.
Scroll horizontally to see all columns →
| Evidence | Possible interpretation | Next step |
|---|---|---|
| Decisions are described as timely. | Decision speed may be a strength. | Keep the useful decision deadlines. |
| Some reports describe little input before decisions. | The consultation process may be unclear or too late. | Invite relevant views before the next planning decision. |
| Different groups describe different experiences. | Context or exposure may explain some differences. | Discuss situations without attempting to identify anonymous raters. |
This is a hypothesis for development, not a diagnosis of Alex’s character. The facilitator should help the participant consider the evidence, identify what is within their control and choose a practical action.
Do not overlook agreement and strengths
If several perspectives describe reliable follow-through, that is useful information even without a gap. The participant can consider how to preserve the strength, apply it in a new responsibility or use it to support a development priority.
Likewise, repeated concerns across groups can justify attention, but repetition does not remove the need to understand situations and examples. Check whether comments describe the same incident rather than several independent observations.
CCL’s interpretation guidance includes strengths, development needs and the relevance of each to future success. See understanding 360 results. A useful plan need not begin with the largest numerical gap.
Protect raters and participants in the actual design
Explain who can see raw responses, individual reports and group summaries. A manager’s feedback may be identifiable by design when there is only one manager. Do not promise that every perspective is anonymous if that is not possible.
Grouping comments does not automatically prevent identification. A distinctive event or phrase may reveal its author. Review disclosure rules, group sizes, combinations of filters and exports. Where necessary, withhold or appropriately summarize detail rather than sharing a traceable quotation.
Preserve enough evidence for authorized quality review without giving every recipient access to every source. The participant’s own report also needs clear handling rules. Establish these arrangements before collecting responses rather than deciding after a sensitive comment arrives.
Use automated analysis as a review aid
Automated theming can help organize comments by behavior and perspective. Test whether it preserves mixed feedback, distinguishes specific examples from general opinions and avoids adding unsupported interpretations.
Do not let sentiment scores stand in for a competency assessment. A warmly worded comment may contain a significant development concern; a blunt comment may still describe a useful strength. Review the substance, not only the tone.
Keep analysis definitions and changes recorded. If one cohort uses “communication” broadly and another separates clarity, listening and updates, comparing theme counts requires care. A shared core can support consistency while allowing role-specific questions.
Turn feedback into a development plan
Choose one or two priorities that matter for the role and are practical to work on. Describe the behavior, situation, support needed and review point. “Communicate better” is difficult to practice; “share decision criteria before the weekly planning discussion” is more concrete.
For Alex, the plan could be to circulate the decision question in advance, invite relevant input and explain the final decision. A later review can ask whether colleagues had a useful opportunity to contribute and whether decisions remained timely.
Do not require a full repeat survey immediately. Give the person a reasonable opportunity to practice. A focused check-in may be useful, depending on the original confidentiality arrangement and the purpose of the follow-up.
Later ratings can inform the review, but changed raters or circumstances may affect them. Document those differences and avoid treating a score increase alone as proof that the plan caused improvement.
Maintain the collection and review process
Sopact’s collection, analysis and governance approach is relevant where teams need to organize recurring evidence and reviewed interpretation. For multi-rater feedback, test the subject, rater-group, period and permission structure carefully. A report about one person can contain information supplied by several others.
Use a pilot with fictional or appropriately authorized data. Include a single-rater group, an identifying comment, missing ratings and contradictory feedback. Confirm what the participant, facilitator and administrator can each see, including in exports or assistant responses.
For the broader method, continue to 360 feedback. For organizational listening requirements, see employee survey software. Keep the development purpose distinct from the company-wide advocacy question described in eNPS.
Watch: reviewing feedback and context
This general feedback video introduces reviewing responses with their context. Multi-rater work additionally needs the rater and participant disclosure arrangements described above.
Frequently asked questions
Is multi-rater feedback the same as 360 feedback?
The terms overlap. Multi-rater means several relevant perspectives; a 360 process commonly draws from several directions around the person. Define the groups and purpose rather than relying on the label.
Are comments more useful than ratings?
They provide different evidence. Ratings can show patterns on defined items, while comments add examples and context. Review them together and consider the quality of both.
Is disagreement the most important finding?
Not always. Differences can reveal useful context, but shared strengths and consistent development needs can be equally important. Prioritize what matters for the role.
Does grouping responses guarantee anonymity?
No. Small groups and identifying details can reveal a rater. Design and test the collection, access and disclosure rules before promising anonymity.
What should happen after the report?
Provide a supported review, choose practical development priorities and agree on a follow-up. A report alone does not ensure that feedback becomes useful action.
Choose your next step
King Center
Related practice: using feedback to support program learning. This story is not a demonstration of multi-rater confidentiality controls.
Read the customer story →