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

# Compare ratings with stats

> Add connected business and activity stats to calibration grids and tables to compare ratings with relevant performance context.

Stats bring additional performance context into a calibration. Alongside manager review ratings, you can use connected data—such as quota attainment or project delivery—as a grid axis or table column.

Stats are evidence for discussion, not editable review answers. A facilitator can change a rating in response to the committee's decision, but cannot change the underlying stat from the calibration.

## How stats work in calibration

Windmill makes eligible stats available for reviewees when the connected data falls within the performance review cycle's review period.

You can use a stat in two places:

* **Grid** — Select a stat as either axis to compare it with a rating or another stat.
* **Table** — Add stats as columns for a compact comparison across the roster.

On the grid, Windmill groups numeric stat values into five peer-relative bands based on the population mean and standard deviation of reviewees who have data in that calibration. This makes it easier to see relative position without changing the source data.

If fewer than five reviewees have a value, or every value is the same, Windmill shows a single band because it cannot create meaningful peer-relative groups.

## Choose useful stats

The strongest stats have a clear connection to the rating being discussed and are reasonably comparable across the selected group.

Useful examples include:

* **Sales:** quota attainment, closed revenue, or pipeline created
* **Customer Success:** retention, expansion, or portfolio health
* **Support:** cases resolved, customer satisfaction, or response time
* **Engineering:** project delivery or relevant quality and reliability measures
* **Recruiting:** hires completed or time to fill

Avoid using activity volume as a shortcut for performance. A high count may reflect role design, assignment mix, or data availability rather than greater impact.

## Build a rating-versus-stat grid

<Steps>
  <Step title="Open Grid">
    Open the calibration and select **Grid**.
  </Step>

  <Step title="Choose the rating axis">
    Select a manager review rating for one axis, such as *Overall performance*.
  </Step>

  <Step title="Choose the stat axis">
    Select a relevant stat for the other axis, such as *Quota attainment*.
  </Step>

  <Step title="Filter the comparison group">
    Narrow the roster to people whose roles and expectations make the comparison meaningful.
  </Step>

  <Step title="Review outliers">
    Open packets where the rating and stat appear misaligned. Look for context before proposing a change.
  </Step>
</Steps>

Because the stat axis is read-only, facilitators can drag reviewees only along the rating axis.

## Example use cases

### Find a rating that needs more context

A salesperson has high quota attainment but a lower overall rating. The committee opens the packet to learn whether the rating reflects collaboration, forecast quality, or another expectation not captured by quota.

### Compare managers' standards

Table view shows rating and outcome data across several teams. One manager's ratings are consistently higher for similar results. The committee reviews packets to determine whether the difference reflects context or inconsistent use of the rubric.

### Avoid penalizing a strong team

A high-performing employee is rated lower than peers on a very strong team. Comparing the broader cohort and opening the evidence helps the committee assess the person against role expectations rather than only their immediate teammates.

## Interpret stats responsibly

* Check whether the stat covers the full review period.
* Compare people with similar responsibilities and opportunity.
* Consider data quality and missing values.
* Use multiple sources of evidence before changing a rating.
* Never treat a stat band as an automatic rating recommendation.

## FAQs

<AccordionGroup>
  <Accordion title="Why don't I see a stat in the axis or column picker?">
    A stat appears only when it is available to Windmill and at least one included reviewee has relevant data during the cycle's review period. Check the integration, stat configuration, roster, and date range.
  </Accordion>

  <Accordion title="Can I edit a stat during calibration?">
    No. Stats are read-only context from connected sources. Update incorrect data in its source system, then allow the integration to sync.
  </Accordion>

  <Accordion title="Can I use two stats on the grid?">
    Yes. Two different stats can form the axes, but both are read-only. Use this view for exploration, then open a packet or switch an axis to a rating if a facilitator needs to propose a change.
  </Accordion>

  <Accordion title="Are stat bands company-wide percentiles?">
    No. The bands are calculated relative to the reviewees with values in the current calibration. Changing the roster can change the bands; filtering only changes which reviewees you currently see.
  </Accordion>
</AccordionGroup>