Skip to content
PEER FEEDBACK

Peer feedback in 360 assessments: the highest-signal rater group

Peers see the assessee doing the work, not presenting it upward. That is why peer data outperforms manager data on collaboration, conflict, and day-to-day execution competencies. The catch: peer rounds also have the highest risk of friendly-fire bias and rater fatigue. This subtopic covers how to pick peer raters, how peer anonymity actually works, and the failure modes that corrupt the peer round.

By Simon CarviPublished April 20267 min read

On this page

Why peer data is the highest-signal rater group

Managers see upward behaviours: the deck, the status update, the conversation the assessee has in front of the manager's peers. Direct reports see downward behaviours: how the assessee manages, delegates, and coaches. Peers see horizontal behaviours: how the assessee collaborates across teams, handles conflict, runs working sessions, shows up in cross-functional meetings. For most mid-career competencies, the peer perspective is where the development priorities actually live.

  • 1Competencies where peers outperform Cross-functional collaboration, conflict resolution, meeting facilitation, information sharing, psychological safety, technical mentoring (peer-to-peer). Manager ratings on these often miss the behaviour entirely.
  • 2Competencies where managers outperform Strategic thinking, executive communication, business acumen, decision quality under ambiguity. Peers often do not see the manager-layer conversations where these play out.
  • 3Convergence is the gold standard When manager, peers, and direct reports all converge on a gap, the development priority is non-negotiable. When they diverge, the conversation has to happen. Either way, the triangulation produces signal.

The 360 rater stack

Five perspectives on the same competencies. Each group sees behaviours the others cannot.

Self

Own view of capability and intent

1 rater

Manager

Upward behaviours: how the assessee presents and decides

1 rater

Peers

Horizontal behaviours: collaboration, conflict, cross-team work

3 raters minimum

Direct reports

Downward behaviours: management, coaching, delegation

3 raters minimum

External

Cross-org view: customers, partners, cross-functional peers

2-3 raters, optional

Groups below 3 raters merge with peers for anonymity. Same competencies, same scale, across every group.

How to pick peer raters

Peer rater selection is where 360 programs quietly fail. Raters the assessee hand-picks with no review produce friendly-fire data. Raters HR assigns in a vacuum produce low-context data. The combination that works is assessee-proposed, manager-reviewed, HR-balanced.

  • 1Assessee proposes 4 to 6 peer names Instruct the assessee to include at least 2 peers they do not naturally work well with. If they only propose friends, the feedback becomes noise. Make the ask explicit: pick peers who will give you honest feedback, not confirming feedback.
  • 2Manager reviews the list The manager cross-checks the proposed list for balance (functions, seniority, relationship mix). They can add or remove but should not replace more than 2 of the original picks. Replacing the whole list signals distrust and the assessee knows.
  • 3HR checks for the minimum-3 rule A peer group of 2 merges with the external group or gets dropped from the report because anonymity cannot be preserved. A peer group of 3 to 6 is the operating range. Above 7, rater fatigue and diminishing returns kick in.
  • 4Exclude direct reports from the peer pool Direct reports have their own rater group (with its own anonymity rules and its own observational bias). Mixing them into the peer pool contaminates both groups. If the assessee does not manage anyone, the direct-report group is simply not used.

How peer anonymity actually works

Peer honesty is load-bearing on the anonymity model. If peers suspect their individual ratings can be traced back to them, they will sanitise. On Huneety, peer anonymity works at two levels: the group average requires at least 3 raters before it displays, and written comments are paraphrased before they reach the report so identifiable phrasing does not leak.

  • 1Minimum 3 raters per group Below 3, Huneety merges the peer group with the external group (if present) or suppresses the group average entirely. The report shows a note: 'insufficient raters for group anonymity'.
  • 2Comments paraphrased by Huna AI Raw peer comments often contain identifying phrasing ('when we worked on the Q3 launch together'). Huna rewrites the comment to preserve the substance while removing the identifier. The assessee sees the point; they do not see who made it.
  • 3Token-based access, no login Raters click a secure link, rate, submit. No password, no account. The token expires after 7 days. No rater identity is stored beyond the invitation record, and the invitation record is separated from the rating record at ingestion.

Peer anonymity built in, not bolted on

Token-based access, minimum-3 group rule, AI-paraphrased comments. Peers answer honestly because the model is trustworthy by design.

See how it works

Three failure modes in the peer round

The same three patterns recur across every first 360 we run with a new customer. Screen for them before launch and the peer round produces signal.

  • 1All peers picked from one team If every peer sits on the assessee's own team, the feedback is in-group biased. The peer set should span at least 2 teams or functions when the assessee works cross-functionally.
  • 2The passive-aggressive peer Occasionally one peer uses the open-text as a performance-review venue. The Huna paraphrase layer softens this, and the debrief should discount clear outliers. If 3 of 5 peers converge and 1 dissents hard, the 3 are the signal.
  • 3Peer round during a crunch Launching the peer round during quarter-end, product launch, or a restructure tanks response rates. Aim for a calm 2-week window. If the calendar refuses to cooperate, push the cycle rather than push through.

Peer feedback on Huneety

Huneety runs the peer round on the same campaign as every other rater group. Invitations go out via token link (no login for peers), reminders fire automatically, responses aggregate by group with the minimum-3 rule enforced, and comments are paraphrased by Huna before reaching the report. HR sees real-time response rates per group. See the Huneety 360 platform.

QUICK ANSWERS

Quick answers

How many peer raters do I need?
3 is the minimum for group anonymity. 4 to 6 is the operating range. Above 7, response rates drop and rater fatigue becomes the bottleneck. Quality beats quantity: 4 peers who see the assessee from different angles beats 8 peers who all sit on the same team.
Who picks the peer raters?
Assessee proposes, manager reviews, HR balances. Each step adds something the other cannot: the assessee knows who has seen their work, the manager knows who will be candid, and HR enforces the anonymity and balance rules across the whole cohort.
Can peer ratings be anonymised fully per individual?
No, and you would not want that. Per-individual anonymity makes the ratings untraceable even to the HR team running the program, which means no way to check for rater fatigue or missing submissions. Group-level anonymity (minimum 3 raters, paraphrased comments) is the right balance.

Launch a 360 with honest peer feedback

Token-based peer access, group-level anonymity, AI-paraphrased comments. The peer round produces signal because the rules make honesty safe.