Without calibration, performance ratings are notoriously inconsistent. Some managers rate every report ‘exceeds expectations’; others reserve top ratings for one in twenty employees. The result is that a high performer in a strict manager’s team receives a worse rating, smaller raise, and weaker promotion case than an average performer in a lenient manager’s team. Over multiple cycles, this drives equity issues, attrition of high performers in strict-manager teams, and inflation in promotion rates from lenient-manager teams.
Calibration sessions typically follow draft ratings being submitted by individual managers. A facilitator (usually an HR business partner or senior leader) leads a group of 6-12 managers through their proposed ratings, asking each to defend any ‘exceeds’ or ‘below expectations’ ratings against specific behavioural evidence. The group challenges, refines, and sometimes revises ratings to match a consistent organisational standard. Final ratings emerge from the calibration session and are then communicated to employees through the performance review meeting.
Key Points: Performance Calibration
- Cross-manager rating discussion: Managers defend proposed ratings against evidence in front of peer managers and a facilitator.
- Consistency over leniency: The goal is shared standards across teams, not forcing a bell curve or stack ranking.
- Evidence-based defense: Calibration only works when managers can cite specific examples and outcomes; vague impressions fail the room.
- Most common before merit and promotion decisions: Calibration typically completes 2-4 weeks before final ratings are communicated and pay decisions are made.
- Effective when paired with rubrics: Calibration without a defined performance rubric devolves into negotiation; calibration with a clear rubric produces consistent results.
How Performance Calibration Works in Treegarden
Performance Calibration in Treegarden
While performance management is typically a separate platform from the ATS, Treegarden’s interview scorecard system applies the same calibration principle: structured rubrics, multiple reviewers, and explicit evidence-citation requirements. The same disciplines that produce well-calibrated interview decisions produce well-calibrated performance ratings.
See how Treegarden handles Performance Calibration → Book a demo
Related HR Glossary Terms
Frequently Asked Questions About Performance Calibration
No, though they are sometimes confused. Forced ranking (also called stack ranking) requires a fixed distribution of ratings - e.g., 20% top, 70% middle, 10% bottom - regardless of actual performance. Calibration aims for consistent standards across managers, accepting that any given team or department may genuinely have above-average or below-average rating distributions. Forced ranking is largely out of favour; calibration is widely practiced.
A typical session covers 30-60 employees in 90-120 minutes. Larger groups quickly become unproductive - facilitators struggle to maintain focus and managers become reluctant to challenge each other. Companies with 1000+ employees typically run multiple parallel calibration sessions organised by function or business unit, then a final senior leadership session that resolves cross-function inconsistency.
Healthy calibration cultures allow disagreement, but the facilitator typically has authority to escalate. Common patterns: (1) require additional evidence within 48 hours; (2) involve the manager’s own manager in the discussion; (3) flag the rating for senior leadership review. The rare case where a manager refuses to budge despite weak evidence is itself useful signal about that manager’s readiness for the rating responsibility.
Most organisations describe the calibration process to employees - typically in the performance management documentation - but don’t share specific session content or who said what. Employees benefit from knowing that ratings are reviewed cross-functionally before being finalised; full transparency on session dynamics would chill manager honesty during the discussion itself.