PayAnalytics March 2026 - Detailed release notes
Release v2026-03-31_01
| Type | Jira | Description |
|---|---|---|
| PA-6511 | A new pay equity target option based on statistical significance was added to the raise suggestion configuration. Compensation managers can now specify a target p-value (between 0.01 and 0.1, defaulting to 0.05) when computing raises to close pay gaps. The UI validates the entered significance level and displays the equivalent standard deviation value to assist with target selection. This feature is currently available behind a development feature flag. | |
| PA-6512 | A post-raise verification method was implemented in the pay gap analysis engine to determine whether the adjusted pay gap after raises meets the significance threshold specified during analysis configuration. The method reuses intermediate regression computation steps to calculate p-values of coefficients after raise suggestions are applied, allowing the system to validate that raise scenarios achieve the intended statistical targets. This calculation is currently accessible programmatically and will support future UI features that display raise effectiveness metrics. | |
| PA-6513 | The analysis configuration view now displays significance-based pay equity targets (p-values) with a 'Significance target' heading and decimal formatting (e.g. 0.05, 0.10) instead of percentage formatting. Analyses configured with traditional percentage-based target gaps continue to display as before with the 'Target pay gap' heading. This change improves the clarity of how different target types are presented after an analysis is run. | |
| PA-7523 | An issue was corrected where the collinearity removal process in pay equity analyses could produce different results depending on the order of rows in the input dataset. When perfectly correlated columns were detected, the system previously dropped columns arbitrarily, sometimes retaining sparse one-hot encoded features that offered little explanatory power. The system now ranks collinear columns by variance and preferentially retains those with distributions closer to 50/50 (high variance), ensuring stable and reproducible regression results. Pay gap analysis reports and regression outputs may show slightly different numerical results compared to before, as different reference groups may now be selected. | |
| PA-7700 | An issue was corrected where the FTE (part-time ratio) field in the dataset configuration step could accept invalid values when set directly via the API, resulting in data discrepancies between the values sent through data integration and those reflected in the system. Server-side validation was added to reject FTE values that are non-numeric, zero, negative, infinite, NaN, or above 5. Users or integrations that attempt to configure a dataset with invalid FTE values will now receive a clear error message. Existing datasets with already-configured FTE fields are not affected. | |
| PA-7739 | An issue was corrected where the subgroup analysis overview page failed to load with a server error when one or more subgroup regressions produced invalid R-squared values. This occurred in analyses with very small sample sizes or single-variable models, where the statistical regression could not compute a valid R-squared metric. The overview page now handles invalid R-squared values gracefully and displays the subgroup analysis results without error. | |
| PA-7751 | An issue was corrected where users with restricted per-record access controls saw action buttons that should have been hidden based on their assigned permissions. The affected buttons included edit and delete options in the analysis selection dropdown and frontpage search results, the Dataset Configuration button, the rename button on dataset search results, and the copy button in the Employee List. Additionally, the Run Analysis button now appears disabled with an explanatory tooltip when users lack the required privileges, rather than being hidden entirely. Users with full record access are unaffected and continue to see all buttons as before. | |
| PA-7758 | An issue was corrected where clicking the Excel export icon on the Salary - Pay Gaps chart in EU Pay Transparency Directive reports did not trigger a download. The issue occurred when charts without underlying summary data were included in the report, causing the export to fail silently. The Excel export now completes successfully and generates a valid file containing available data from the report. | |
| PA-7773 | Reports dropdown menu is cut off by the top of the browser viewport when page is scrolled to the top | |
| PA-7825 | An issue was corrected where the Median hourly salary pay gap in the EUPTD report displayed "negative infinity %" or blank values when the reference group had no salary data, caused by division by zero in the underlying calculation. The report now displays a dash ("–") for any formula-based metric that results in an undefined numeric value. | |
| PA-7852 | An issue was corrected in the UK Gender Pay Gap, Irish Gender Pay Gap, and EU Directive Pay Gap reports where the proportion of men and women receiving a bonus payment was incorrectly calculated. The calculation divided the number of employees receiving a bonus by the total employee count rather than by the specific gender group total, producing artificially low percentages. The formula now correctly divides by the gender-specific headcount, ensuring accurate bonus proportion reporting for government-mandated pay gap reports. | |
| PA-7857 | An issue was corrected where tenant creation in the development environment failed with an out-of-memory error. The container running the tenant creation process was configured with insufficient memory (512 MB) to accommodate both the listener process and the Django subprocess that executes database schema migrations. The memory allocation was increased to prevent the failure. |
Release v2026-03-26_01
| Type | Jira | Description |
|---|---|---|
| PA-7845 | An issue was corrected where the Pay Equity Tool was inaccessible for instances configured with CSRF exemption, despite the instance showing as active in the administration panel. This was caused by a CSRF validation error. The tool now loads correctly for CSRF-exempt customer configurations. |
Release v2026-03-24_01
| Type | Jira | Description |
|---|---|---|
| PA-7790 | An issue was corrected where p-values in pay gap analyses were computed using a normal (z) distribution approximation instead of the statistically correct t-distribution. This particularly affected analyses with smaller groups, where the lower degrees of freedom require the t-distribution's thicker tails for accurate significance testing. The calculation now uses the Apache Commons Math t-distribution method, producing correct p-values across all group sizes and eliminating floating-point precision errors that previously caused p-values to bottom out at approximately 1e-15 for large test statistics. | |
| PA-7801 | An issue was corrected where the button label for excluding employee groups from raise suggestions was cut off when the platform interface was displayed in German. The button text now wraps correctly and displays in full. |
Release v2026-03-17_01
| Type | Jira | Description |
|---|---|---|
| PA-7657 | A new "Raise suggestion configuration" section was added to the System Parameters page, replacing the previous "Raise suggestions" section. Three new configuration settings were introduced: Minimum eligible size for a reference group (default: 7), Minimum % relative to total population (default: 10%), and Minimum % relative to largest demographic group (default: 25%). HR admins can now configure these thresholds to control which reference groups qualify for raise suggestion calculations based on population size criteria. | |
| PA-7658 | The reference group eligibility settings used for an analysis are now displayed in the Analysis Configuration pop-up. | |
| PA-7708 | An issue was corrected where compensation after raise was incorrectly calculated when using compensation component aggregates that included scaled compensation components. The calculation incorrectly applied FTE scaling to the raise amount, resulting in inflated post-raise compensation totals. Compensation after raise now correctly reflects the unscaled sum of raw compensation components plus the raise amount. | |
| PA-7732 | An issue was corrected where the raise adjustment dialog displayed incorrect values for part-time employees. When changing a raise amount, the dialog showed the scaled compensation value instead of the actual employee compensation in the 'Compensation after raise' field, and the raise percentage did not update correctly. The dialog now displays the correct actual compensation values and updates the percentage appropriately for part-time employees. | |
| PA-7733 | An issue was corrected where raise percentages were incorrectly displayed and applied for part-time employees when scaling was enabled. When editing a single employee's raise or applying bulk raise updates with a fixed percentage, part-time employees would show inflated raise percentages in the employee table (for example, 20% displayed instead of the entered 10% for employees in 50% positions). Raise percentages now display and calculate correctly for part-time employees in both individual and bulk raise workflows. |
Release v2026-03-10_01
| Type | Jira | Description |
|---|---|---|
| PA-7644 | A "Category of Worker" field selector was added to the dataset configuration page in the ID and Group section. Users can now designate an existing dataset field as the Category of Worker, which then prepopulates as the default selection in the Close Pay Gaps by Groups feature, the EU Directive Report wizard, and Pay Transparency Reports. This field supports EUPTD compliance requirements for worker category segmentation. When the Category of Worker field is cleared, all three features revert to using the main employee characterization field as the default. | |
| PA-7662 | Pre-population logic was added for the Category of Worker field across analysis and reporting features. When a Category of Worker field is defined in the dataset configuration, it now automatically populates as the default selection in the Close Pay Gaps by Groups analysis, the EU Pay Gap Report wizard, and Pay Transparency Reports (including Employee Right to Request and Pay Explainability reports). Users retain the ability to override the pre-populated selection in any feature, and changes to the Category of Worker field after analysis has been run affect only future report generation, not historical results. | |
| PA-7705 | An issue was corrected where the Results section did not appear in the Analysis Overview tab when running subgroup analyses with a non-gender demographic (such as Race, Age, or Location) configured as the first demographic. The Results section now displays correctly with pay gap metrics regardless of demographic ordering. | |
| PA-7765 | Automatic daily cleanup of historical uploaded datasets based on each organization's data retention policy was implemented. Organizations with data retention policies enabled will now have their uploaded datasets automatically removed once the uploaded data exceeds the configured retention period (minimum 365 days). The retention period is calculated from the uploaded_time of the dataset, not the effective_date of the pay data. Organizations without a retention policy enabled are unaffected. |
Release v2026-03-03_01
| Type | Jira | Description |
|---|---|---|
| PA-7709 | Reference group eligibility settings were exposed via the settings API and analysis configuration endpoint. HR admins can now view and configure thresholds that control which employee groups qualify for inclusion in pay equity analysis, including minimum population sizes and percentage thresholds. These settings are editable through the settings interface and displayed (read-only) in the analysis configuration view, allowing environment-specific tuning of reference group criteria. | |
| PA-7738 | The Pay Equity Overview report was made the default report when an analysis is selected, replacing the Compensation Distributions by Demographic report. | |
| PA-7761 | The Standard Deviation column was removed from tables in the Compensation Assistant and Pay Explainability pages. This change streamlines the tables to display only metrics that directly relate to the visualizations shown in the graphs above each table. |