Data column terminology & definitions
As a result of running a full pay equity analysis, including measuring pay gaps and suggesting raises, a number of data columns are added to your employee dataset. These columns provide employee level pay equity insights to help you make informed decisions on the final raises granted to each employee.
In this article we will cover three topics:
Where to find, view and remove columns from displaying in your employee list.
Best practice approach using the columns to make informed decisions
Finding, viewing & removing pay equity columns
The columns will only become available within your employee dataset after running a full pay equity analysis. To find, view, and remove the additional pay equity columns, proceed as follows:
From the navigation menu select Pay Equity Analysis.
In the Pay Equity Analysis page scroll down to the Employee list to view your data.
Use the horizontal scroll bar within the employee list to view the final columns.
You will already see the basic columns, starting from Predicted compensation. Note that not all columns will automatically be displayed
To view more columns, click Table columns and select the additional columns you would like to display in the employee list.
Similarly, you can unselect columns that you would like to remove from the display.
The Employee list section is illustrated in the following picture:
The Employee list
Column definitions, calculations & relevance
After running your full pay equity analysis, data columns will be added to your employee data in the Employee list. In this article, we divide the additional column data into two types, which we’ll cover separately below. Each data column type covers a different stage of the analysis:
Compensation model columns - Columns that focus on the difference between actual and predicted compensation based on the configured and computed compensation model such as predicted compensation and outlier percentages.
Raise suggestions columns - Columns that focus on the compensation adjustments to achieve set pay equity goals within requirements and boundaries such as raise suggestions, and the raise- floor and ceiling.
Compensation model columns
The columns display various results from the pay equity analysis to identify pay gaps. Each column provides a different insight into indicators of pay gaps based on the current compensation and the predicted compensation from the regression model.
Predicted Compensation: Predicted compensation, refers to the estimated pay that an employee should receive based on the objective factors used in the analysis, such as their job position, experience, location, and other relevant variables. That is, the predicted compensation is determined on an individual employee basis by accounting for numerous factors related to aspects of the job (such as the role or the responsibility) and factors related to the employees themselves (for example experience, tenure, or even performance). Regression analysis allows us to account for all of these aspects together when running the analysis and calculating the equal pay gap. That means the predicted compensation is the salary an employee should receive based on the input factors used in the analysis, only using internal data. For further information on how the system calculates the predicted compensation, check the Frequently asked questions page.
Distance from Predicted Compensation: The difference between an employee's actual compensation and their predicted compensation, expressed as a number, and computed as actual compensation minus predicted compensation. An employee with actual compensation below their predicted compensation is represented by a negative number.
Outlier Percentage (%): The percentage by which an employee's actual compensation deviates from their predicted compensation, indicating whether their pay is (unexpectedly) high, indicated by a positive percentage, or low, indicated by a negative percentage. An employee is considered to be ‘an outlier’ if their actual compensation deviates from their predicted compensation, can either be a positive or a negative outlier. When entering the “outlier mode” you can set the filter to a set outlier %, for example, 30%, and analyze all employees who have salaries that deviate more than 30% from the predicted salary.
Outlier (Standard Deviation): A measure of how far an employee’s actual compensation is from the predicted compensation, expressed in terms of standard deviations from the mean (or predicted compensation). When entering the “outlier mode” you can set the filter to a set outlier standard deviations, for example 2, and analyze all employees who have salaries that deviate more than 2 standard deviations from the predicted salary.
Compensation adjustments (raise analysis):
The columns display various results from the pay equity analysis to inform raise decisions. Each column provides a different insight into the parameters for setting pay raises and the impact of suggested and actual pay raises awarded.
Compensation Floor: When pay bands are factored into the analysis to address the pay gap, the system will display the minimum salary for each employee based on their respective pay band. This represents the lowest permissible compensation for an employee according to the applicable pay band.
Compensation Ceiling: When pay bands are included in the analysis, the system will display the maximum allowable salary for each employee based on their respective pay band. This represents the highest permissible compensation for an employee according to the applicable pay bands.
Raise Ceiling: The maximum allowable raise amount applied to the employee's compensation. The raise ceiling can be based on two factors, the maximum raise configuration and/or pay bands.
The maximum raise for each employee is based on the maximum raise % set in the analysis configuration, multiplied by the actual compensation. The second factor is pay bands. In this case, the raise ceiling is the amount equivalent to the difference from the current salary to the compensation ceiling. This refers to the salary increase an employee can receive while remaining within the limits of their designated pay band.
Diff. from Compensation Floor before Raise: The amount difference between the employee's current compensation, before any raise, and the compensation floor. In exceptional cases where the employee’s compensation is below the compensation floor, the amount will be negative. Generally, the amount will be positive. Together with the diff. from the compensation ceiling before the raise (see below), this measure helps identify the employee's position within the boundaries of their pay band.
Diff. from Compensation Ceiling before Raise: The amount difference between the employee's compensation before raise, and the compensation ceiling.
Suggested Raise: The amount of salary increase suggested for an employee by the system based on the raise suggestion configurations in the pay equity analysis and the applicable pay gaps from the compensation model. The computation of suggested raise is complex and based on a PayAnalytics internal model, incorporating user settings such as pay equity targets, group vs. individual fairness, cost effectiveness vs. fairness, and restrictions such as maximum and minimum raise % among others. The suggested raise amount is by default also populated as the Raise amount (see below). However, the suggested raise is exactly that, a suggestion. It is up to the organization to decide whether to apply the suggestion or deviate. If all suggested raises are applied as actual raises to employees, the organization will achieve set pay gap targets.
Raise: The actual raise amount applied to the employee. By default, the raise amount is equal to the suggested raise amount. This value can be edited by the user. After editing the raise amount, the pay gaps may be impacted, allowing you to view the impact of deviating from suggested raise suggestions at an employee level. At any point in time, the edited raises can be reset back to the PayAnalytics suggested raises. To edit raises, click on Update Raises above the employee list, as illustrated on the following picture:
Update raises pop-up
From there, you can set all raises to zero or apply the original PayAnalytics suggestions. You can also choose to set raises as a fixed amount or percentage, or even add fixed amounts on top of existing raises for individual employees. Additionally, you can use the filter feature in the employee list to apply updates to a specific subset of employees. For example, if you want to apply a fixed percentage raise to all unionized employees, simply use the filter and update the raises accordingly. Lastly, you have the option to adjust raises so that all employees are above the minimum of their respective pay bands.
Raise (%): The percentage increase in compensation after applying the employee level raise to their actual compensation. Raise (%) is computed as the actual raise divided by current compensation. In case a maximum raise (%) was set in the raise configuration step, the initial raise (%) based on suggested raises should never exceed the maximum raise (%). However, one exception that can cause the raise percentage to exceed the maximum allowable is if an outlier threshold is applied in the raise configuration. For example, if the system is set to bring employees below a certain threshold—such as 20%—up to that level, it will suggest raises to all employees, regardless of their demographic value, to ensure they meet or exceed the threshold. As the raise amount is edited from the suggested raise, the raise percentage may exceed the maximum. The raise percentage can also help to identify extreme cases, either where (many) relatively small raises are awarded or on the other extreme, where (some) extremely high raises are awarded. Both could have implications for policy and/or operational implementation.
Outlier Percentage After Raise (%): The percentage deviation of an employee's new compensation, after the raise, from their predicted compensation. In most cases, where a raise is applied, one would expect the outlier percentage after raise (%) to reduce. However, in specific pay gap scenarios and raise configuration settings, this may not hold true. For example, when cost-effectiveness is prioritized over fairness and group is prioritized over individual fairness, raises might be awarded to employees, even when their compensation already exceeds their predicted pay. To reduce, or even avoid, these raise suggestions, prioritize individual fairness.
Raise Effect on Pay Gap: The estimated impact of the raise on closing the overall existing adjusted pay gap. The cumulative “raise effect on pay gap” across all employees is equal to the difference between the adjusted pay gap before and after raises. In some cases, particularly in datasets with a high volume of employees, the raise effect on pay gap might be 0.00 despite a raise. One might conclude that the raise does not have any impact on the pay gap. Due to two decimal rounding, the effect appears to be zero. However the individual effect may be small, the cumulative of many employees can still have a significant impact on pay gaps.
Similarity Score: A score (ranging from 0 to 100) that reflects the similarities between employees. The similarity score is generated only when the “Find Similar Employees” feature is used. For more detailed information on how to use the feature, see Finding similar employees .
Best practice approach using the columns to make informed raise decisions
By default the PayAnalytics platform displays the following result columns:
Predicted compensation
Outlier percentage (%)
Sugg. raise
Raise
Raise (%)
Raise effect on pay gap
Compensation after raise
In order to finalize your employee level raises and achieve your pay equity goals, we suggest as a best practice to review the additional compensation and raise columns in the following order:
Outlier percentage (%): Explore extreme cases to establish whether these are valid or whether explanations can be found that were not considered in the initial pay equity analysis. Maybe you need to update any data fields or include additional objective factors in the compensation model to better capture the pay variability.
Raise (%): Review the extreme cases, both low and high, and assess the overall distribution of raises to ensure they align with your policy and operational goals. For example, do you want to approve raises exceeding 30%, or even 100%, in a single instance, or would you prefer to phase these raises over multiple periods? If there are many small raises, consider whether the administrative cost is justified by the impact, or if it would be better to set a minimum raise threshold. This could reduce the number of raises while enabling you to award fewer, but higher, raises to achieve the pay equity target?
Raise effect on pay gap: Consider the adjusted pay gap in your analysis results and the pay gap target for the current round of raise suggestions. Do you plan to close the pay gap entirely (to 0%), or would you prefer to take a more gradual approach, reducing the gap in smaller, incremental steps?
Raise: Where applicable, edit and finalize the raise amounts and export the file to make relevant updates in your compensation (and payroll) system(s).
There is no one-size-fits-all when it comes to pay equity and depending on the specific circumstances and complexity of your organization you may want to consider reviewing different columns. PayAnalytics was built to keep things simple initially, with the option of adding in further information if required.