Skip to main content
Pay equity analysis

Insights and compensation outliers

Dig into pay bands, gap drivers, demographics and employees paid far from their peers.

Written By Karl

Last updated About 2 hours ago

The Insights and Comp Outliers tabs help you find what is behind your pay gaps: which employees drive them, how pay sits against bands, and who is paid well above or below comparable colleagues.

Where: Pay Equity → Insights and Pay Equity → Comp Outliers

Insights

  • Pay Bands: where employees sit in their bands. Choose Group by, and under Show pick Defined pay bands, Current salary percentiles or Compa ratio. The shaded 80–120% area is the target range.

  • Compa Ratio: how pay sits against band midpoints. It stays empty until employees match a pay band.

  • Pay Gap Drivers: the employees whose pay moves the organisation gap the most, ranked High impact, Moderate or Low. Use Exclude from analysis to see the result without someone, and Include in analysis to undo it.

  • Employee Insights: compare pay between keyword groups, such as high performers or recent hires. See Keywords.

  • Age Group Analysis, Demographic Breakdowns (quartiles by level, grade, family and more), Performance vs Compensation and Most Valuable Skills.

Comp Outliers

  • ML Predicted vs Actual Compensation: each employee's actual pay against the pay a regression model predicts from their attributes. The model doesn't use gender. Select Show model performance for the model's statistics.

  • Outlier Analysis: employees paid more than 20% above or below their position's median, split into Overpaid and Underpaid and sorted by impact.

To document why someone is an outlier, select Justify. Each row shows how much is Explained, Remaining and Justified.

Interpreting predictions and outliers

Predicted pay is an analytical estimate based on the model and the data available to it. It is not a recommended salary, a finding that someone's pay is fair, or a legal conclusion. Treat an outlier as a prompt to check the source data, comparison group and documented reasons for the difference before making a decision.

Good to know

  • Outlier analysis needs employees to have positions.

  • The predicted pay chart uses current data, so it isn't shown when you view a past analysis run.

  • Excluding a driver only changes your view of the analysis, not the employee's data.

Was this helpful?

Still need help? Share an idea