Peer feedback becomes unreliable the moment it is treated as an informal popularity exercise. Knowing how to collect peer feedback means designing a process that captures relevant evidence from the right people, protects candour, and gives managers information they can use in a calibrated performance discussion.

For HR teams and managers, the objective is not to collect more commentary. It is to collect comparable, role-relevant input that helps distinguish sustained performance from isolated impressions, visibility from contribution, and development needs from personality preferences.

Start with the decision the feedback must support

Peer feedback should have a defined operational purpose before anyone receives a questionnaire. Is it intended to inform a performance review, identify development priorities, test cross-functional collaboration, or support a promotion decision? The purpose determines the questions, respondent group, confidentiality settings, and level of evidence required.

A broad request such as “What do you think of working with this person?” creates broad and inconsistent answers. One reviewer may comment on attitude, another on technical quality, and a third on a single difficult meeting. The result is difficult to compare and even harder to defend.

Instead, connect feedback to the person’s actual role responsibilities. A project lead may need input on planning, stakeholder communication, risk management and delivery quality. A people manager may need evidence on delegation, coaching, decisions and team accountability. When questions reflect the work the employee is expected to perform, reviewers have a clearer basis for judgement.

This is where role analysis matters. Position descriptions often contain generic language, while the day-to-day job has specific tasks and accountabilities. Translate the role into observable responsibilities before building the feedback process. AI-assisted task extraction can reduce the administrative effort, but HR should still validate the output with the relevant manager.

Select peers who have enough evidence

The right peer is not necessarily the most senior person, the closest colleague, or the person with the strongest opinion. Select people who have observed the employee’s work closely enough to comment on the criteria being assessed.

A practical respondent group often includes direct collaborators, internal stakeholders, and cross-functional partners. The number depends on the role and organisation, but a smaller, well-chosen group is usually more useful than a large group with limited exposure. For many roles, three to six informed peers produces enough perspective without turning the process into a survey campaign.

Ask managers to nominate reviewers, then apply HR controls. Left entirely to the employee, peer selection can become overly favourable. Left entirely to the manager, it can reflect one manager’s network or assumptions. A combined approach is more balanced: employees can suggest people with direct knowledge of their work, managers can add required stakeholders, and HR can check for gaps or conflicts.

Avoid requesting feedback from people who have had little meaningful contact with the employee. Their responses tend to rely on reputation, office visibility or hearsay. That can amplify bias rather than improve the quality of the review.

Use questions that ask for observable evidence

Good peer feedback prompts reviewers to describe work, outcomes and behaviours. Poor prompts invite labels. “Is this person a strong team player?” encourages vague ratings. “How effectively does this person share information needed for your team to deliver its work?” gives the reviewer a clearer frame.

Use a consistent rating scale across respondents, then pair it with a short evidence field. Ratings make patterns easier to identify; comments explain the pattern. Neither is sufficient on its own.

For each responsibility, ask questions that cover a practical mix of delivery and collaboration:

• What did the employee do that helped or hindered the work?
• What was the impact on delivery, quality, customers or other teams?
• How consistently did this occur during the review period?
• What is one specific action that would improve their effectiveness?

Keep the questionnaire focused. Asking peers to assess every capability in a competency framework leads to rushed answers and generic comments. Limit the assessment to the responsibilities peers can reasonably observe. A finance partner may be able to assess a manager’s planning and commercial judgement, but not their day-to-day coaching of direct reports.

Protect confidentiality without removing accountability

Confidentiality is often necessary for candid peer feedback, particularly where there are power dynamics or difficult working relationships. But confidential should not mean unstructured, anonymous criticism with no quality control.

Set clear expectations before the process opens. Explain who can see individual responses, whether comments will be aggregated, how feedback will be used, and what behaviour is unacceptable. Personal attacks, speculation, discriminatory language and feedback unrelated to the role should be excluded.

For small teams, anonymity may be impossible in practice. A detailed comment can reveal its source even if names are removed. In that situation, use aggregated themes, broaden the reviewer group where appropriate, or ask a trained HR partner to synthesise the input. The right approach depends on team size, the sensitivity of the review, and local employee relations requirements.

Confidentiality also needs a deadline and a clear submission process. Open-ended requests produce late responses and reduce comparability. Give reviewers enough time to provide considered input, usually five to ten working days, and send one purposeful reminder rather than repeated chasers.

Separate peer input from manager influence

One of the most common weaknesses in feedback collection is premature influence. If peers know the manager’s view or can see another reviewer’s comments before submitting their own, their responses can shift towards the dominant narrative. That is especially likely when the manager is highly visible or the team has an established view of the employee.

Collect responses independently and lock submissions once completed. For performance reviews, the same principle should apply to self-assessments and manager assessments. Each party should submit their evaluation before seeing the other’s responses.

A blind, side-by-side comparison model creates a cleaner discussion later. Differences do not automatically mean someone is wrong. They identify where the manager, employee and peers have interpreted performance differently, and where the discussion should focus on evidence. CALIBRATE applies this logic by keeping assessments locked until both parties submit, reducing the opportunity for early feedback to shape the outcome.

Analyse patterns, not isolated comments

A single negative comment can be valuable, particularly if it identifies a material risk. It should not, however, be treated as a verdict without corroboration. Review peer feedback for consistency across respondents, relevance to the role, recency, and the specificity of the evidence provided.

Look for patterns such as repeated delays in handovers, strong recognition for technical quality paired with concerns about stakeholder updates, or different experiences across teams. These patterns are more useful than an average score alone because they point to conditions that may be affecting performance.

Managers should also compare peer feedback with delivery data, customer outcomes, self-assessment, prior goals and their own direct observations. Peer feedback is an input to judgement, not a substitute for management judgement. It can reveal blind spots, but it can also reflect uneven exposure, interpersonal friction or a team’s preference for a particular working style.

Where accounts conflict, investigate rather than average them away. Ask whether the employee performs differently under pressure, with certain stakeholders, or in different parts of the role. That is a more productive calibration question than deciding whose opinion carries the most weight.

Turn feedback into a focused discussion

Feedback collection fails when responses disappear into a review file. Managers need to translate the evidence into a fair conversation about expectations, demonstrated performance and next actions.

Share themes rather than reading every comment aloud. Be precise about what the evidence suggests: “Several project partners noted that your work is accurate and dependable, but they often receive status updates too late to adjust their own plans.” This is more useful than saying the employee needs to communicate better.

Agree on one or two development actions linked to the role. If communication is the issue, define the required behaviour, audience and timing. For example, a project lead may commit to a weekly risk update, clear decision records, and escalation within an agreed timeframe. Progress can then be reviewed against observable evidence rather than a vague intention to improve.

The strongest peer feedback process gives people a fair opportunity to be assessed on the work they actually do. When criteria are clear, reviewers are well chosen, submissions are independent and managers examine evidence carefully, feedback becomes a calibration tool rather than another source of review noise.