Performance Calibration

Individual managers see different parts of a business, so one may call a result excellent while another calls similar work satisfactory. Calibration puts those assessments side by side.

Agree on purpose first, because a discussion for development may differ from one used for bonus or promotion decisions, and employees should know the broad process. Define the review period, since mixing a full year's work with the most recent project can create unfair comparisons, and set criteria before names are discussed.

Goals, role expectations, behaviour and relevant results should be explicit, and nobody should invent a new standard for one person mid-meeting. Prepare evidence, with managers bringing examples of work, outcomes and feedback, and context about the employee's opportunity to perform.

Distinguish outcomes from circumstances, because a team with better resources or easier accounts may produce higher raw numbers, and check role differences, since a project lead, customer agent and engineer may need different measures. Calibration seeks consistent application of relevant standards, not identical numerical targets.

Use a facilitator who asks for evidence, watches time and challenges unsupported claims, because seniority should not decide the rating by itself, and give each person sufficient attention so that a rushed review of the last few names does not make meeting order matter more than performance. Question recent-event bias, as one high-profile success or mistake may overwhelm months of other work, and question proximity bias, because staff who work remotely or have less access to leaders may be less visible.

Avoid forced distribution unless a justified policy requires it, since a predetermined number of low or high ratings can override actual evidence. Record reasons for changes by saying what new fact or shared standard supports the move, because "the room felt different" is not enough, and protect dissent so that a manager can explain a well-supported view even if others disagree, since groupthink can create fresh bias.

Check patterns by reviewing rating changes across departments and relevant protected groups under local law, with privacy safeguards and appropriate HR expertise, and keep the discussion on job-related evidence and criteria rather than a personality contest. Explain outcomes to employees appropriately, because a useful performance conversation names strengths, gaps and development steps while a bare score offers little direction.

Separate pay decisions where useful, since compensation budgets can pressure a meeting to change ratings, and if ratings and pay are linked, make that relationship clear; keep records confidential, reaching only authorised people and retained under local rules. An adjustment rate can flag process issues: if 18 of 120 ratings change, 15% were adjusted, though a high or low rate alone does not prove fairness.

Review managers' standards over time, since repeated inflation or downward changes may show a need for training or clearer role criteria, and invite follow-up evidence when appropriate because a decision need not be rushed to fill a dashboard. SHRM warns that calibration discussions can reinforce groupthink, favouritism and centrality bias, CIPD's performance-management review stresses fair and useful judgments and the distinction between developmental and administrative appraisals, and for an owner the test is whether similar evidence receives similar treatment and people get meaningful feedback, not whether all teams have neat rating curves.