> ## Documentation Index
> Fetch the complete documentation index at: https://docs.elevate.help/llms.txt
> Use this file to discover all available pages before exploring further.

# Calibration

> How the Performance Team runs calibration—A-player cap, missing data, outliers—and how leadership approves final grades.

Calibration is the set of **checks and reviews** the [Performance Team](/process/roles-and-ownership#performance-team) runs in **Step 2** of the [Quarterly Cycle](/process/quarterly-cycle). It happens after grades are calculated from [Scorecards](/framework/scorecards) and the [Talent Bar](/framework/talent-bar), and before results are announced.

## Purpose

Calibration ensures:

* **Consistency** — Grades align with our [High-Performance Model](/philosophy/high-performance-model) (A-players \~15–25%, underperformers appropriately identified)
* **Defensibility** — No unexplained outliers or missing evidence
* **Fairness** — Similar performance gets similar grades across managers and teams

## Calibration checks

### a) A-player cap (\~15–25%)

* A-players should account for **no more than \~15–25%** of the population.
* Prevents rating inflation and overly generous managers.
* If the share is too high, the performance team and leadership review and adjust (e.g. recalibrate borderline cases).

### b) Missing data or insufficient justification

* Every [grade](/framework/performance-grades) should have **completed [scorecards](/framework/scorecards)** and enough [evidence](/platform/connectors-and-evidence).
* Gaps are flagged; [managers](/process/roles-and-ownership#managers) or the [performance team](/process/roles-and-ownership#performance-team) fill them before results are final.
* Ensures outcomes can be explained to employees and to leadership.

### c) Outliers

* **Sudden changes** — Big swings from prior quarters without clear explanation.
* **Extraordinarily high or low** — Grades that don’t match the written evidence or that skew the distribution.

Outliers are reviewed. They may be corrected, or the evidence may be confirmed and the grade kept.

## Leadership review

* **[CEO and top management](/process/roles-and-ownership#ceo--top-management)** review and approve final [grades](/framework/performance-grades) before Step 3.
* They check:
  * Fairness and consistency across the org
  * Areas needing intervention (e.g. teams with many underperformers, or persistent calibration issues)
* Performance is a **direct CEO mandate**; this review is how that is applied operationally.

## Output

After calibration and leadership approval:

* **Final calibrated [grades](/framework/performance-grades)** are locked.
* The [Performance Team](/process/roles-and-ownership#performance-team) coordinates the **results announcement**.
* [Managers](/process/roles-and-ownership#managers) deliver feedback and [Outcomes](/process/outcomes) in Step 3.

Calibration is what turns raw scorecard data into **trusted, actionable** [Performance Grades](/framework/performance-grades).