What is Rating Normalization? Meaning and Example

Why ratings need normalizing

Two managers seldom rate the same way. In a retail chain, the Pune regional manager may rate most store managers 4 out of 5, while the Nagpur regional manager rarely gives above 3. If increments follow raw ratings, Pune store managers earn more for the same performance. Normalization looks for such patterns: the average rating per manager, the spread of ratings, and whether a team's ratings match its business results. It then adjusts ratings so that a 4 means roughly the same thing in every region.

Common normalization methods

In practice, companies combine these methods. A statistical check spots managers whose ratings sit far above or below the rest, and a calibration discussion then decides what to change and records why each rating moved.

• Forced or guided distribution: fit each group to a target spread, as in the bell curve
• Calibration meetings: managers review each other's ratings with evidence and agree changes
• Statistical adjustment: scale each manager's ratings to a common average and spread
• Anchoring to results: compare team ratings with the team's actual KPI outcomes

Explaining normalized ratings to employees

Normalization is where appraisals lose trust. An employee told '4' by the manager in March who then sees a '3' in the increment letter will feel cheated. Avoid this by having managers share ratings only after normalization, explaining the method before the cycle starts, and giving the reason for any change in the appraisal discussion. Keep a record of the original rating, the final rating and who approved the change.

Frequently asked questions