PayAnalytics - September 2026 detailed release notes
Release v2026-09-29_01
| Type | Jira | Description |
|---|---|---|
| PA-8574 | When setting up a Pay Transparency bulk export, users now choose which compensation components the generated reports cover, matching the choice already offered for single-employee reports. The selection opens on the user's last choice for that dataset, and both flows explain that the components on offer come from the dataset configuration. | |
| PA-8575 | Reports generated by a Pay Transparency bulk export can now include an Employee details sheet, laid out as in single-employee reports. A new Employee information picker in the bulk export modal sets which dataset columns the sheet lists, opening on the user's last choice for that dataset and with nothing selected otherwise. | |
| PA-8627 | The suggested compensation in the Compensation Assistant could differ from the analysis's predicted compensation for the same employee when a categorical variable held a whole-number value in a column stored as decimals. The Compensation Assistant now reads such values as intended, and predicted compensation figures stored by earlier runs are corrected when the analysis is run again. | |
| PA-8648 | When remediating pay gaps per category (as required under the EU Pay Transparency Directive) with the Group Fairness slider set above 0%, raises are weighted toward the employees who fall furthest below the pay line of their category's reference gender (e.g., men or women). With this fix, the group fairness ranking is measured against each category's own reference gender rather than the analysis-level one. Previously, in categories remediated against a different reference gender, this produced a flat ranking and raises were spread evenly across employees. Raises are now targeted as intended. For more on how group fairness works, see Getting suggested remediation actions. | |
| PA-8796 | The Regression results table on an analysis's Compensation Model tab now shows each variable's SD (standard error) and T-statistic, and the Download Excel file includes both. Users can see how precisely each coefficient is estimated. Analyses run before early August 2026 show dashes in these columns until they are run again. | |
| PA-8812 | A change made to a report right after opening it, such as selecting a new variable for all of its sections, could be discarded without warning, because opening the report also saved it and that save overwrote the change. Opening a report no longer saves it when nothing has changed, so changes made straight away are kept. |
Release v2026-09-22_01
| Type | Jira | Description |
|---|---|---|
| PA-8071 | In datasets that scale compensation to full-time equivalent and include part-time employees, the adjusted pay gap could differ between the before-raises and after-raises views even when no raises were applied. Pay gaps and group averages are now based on full-time equivalent pay, so both views match when no raises are given. Generated reports reflect the same figures. | |
| PA-8555 | The Compensation Model tab of a Pay Equity Analysis now explains when a negative impact on compensation is caused by overlapping compensation drivers (multicollinearity, where two or more drivers carry largely the same information and the model splits their shared effect). A warning names the overlapping drivers, shows how much they overlap, and recommends removing one of them or refining their definitions. This helps HR analysts tell a modeling effect apart from a data issue. | |
| PA-8723 | Uploading a large employee dataset could fail with an error saying the file had no column header, even though a valid header was present. Large datasets now import correctly. Upload error messages are also clearer: header errors name the affected columns, and CSV files not saved in UTF-8 encoding are now identified as such, rather than reported as an unsupported file type. |
Release v2026-09-16_01
| Type | Jira | Description |
|---|---|---|
| PA-8733 | Large employee-data uploads to the public API, whether sent as a single request or in chunks, could fail with an unexplained error. Both upload methods now accept the previously supported size again, and a request that is genuinely too large returns a clear error message. |
Release v2026-09-15_01
| Type | Jira | Description |
|---|---|---|
| PA-8591 | Downloading an Excel file from the Pay Gaps by Category view exported every column instead of only the ones displayed on screen, unlike other Excel exports in the platform. It now includes only the columns currently selected for display, matching how the rest of the platform behaves. | |
| PA-8612 | The pay equity target step of the analysis setup now offers a significance goal alongside target pay gap and target budget, with the threshold set as a number of standard deviations or as a significance level. Both units stay visible together, and the saved-configuration summary on the pay adjustments screen reflects whichever one was set. This capability is currently available to a limited number of customers, with broader availability planned. | |
| PA-8722 | On the analysis overview, the significance verdict on the adjusted pay gap after raises tile now clears once the raises on screen differ from the run's suggestion – the same behavior the Pay Gaps view already has. The verdict returns when the raises are reset to match the suggestion. This capability is currently available to a limited number of customers, with broader availability planned. | |
| PA-8733 | Sending employee data to the public API in a single request could fail unexpectedly once the request grew beyond a few megabytes. Larger single-request uploads are now accepted, and a request that is genuinely too large returns a clear error explaining why. | |
| PA-8745 | Previously, the language picker and language column in the Pay Transparency bulk export feature listed options by their native names, regardless of the platform's settings. Both now match the current UI language. | |
| PA-8751 | So far, the Pay Transparency bulk exports page and its sidebar entry have used different names for the same feature. Both now consistently read "Pay Transparency Bulk Exports." | |
| PA-8762 | Previously, the analysis overview's significance verdict appeared only when comparing a single demographic variable with two levels. It now appears for every analysis run against a significance goal, including those with more levels or more than one demographic variable. Also, the results table gains verdict, significance level, and standard-deviation columns for each targeted group. This capability is currently available to a limited number of customers, with broader availability planned. | |
| PA-8766 | Until now, the statistical significance verdict on the Pay Gaps tab and the analysis overview could appear even when the reference group on screen was not the one the analysis actually measured the pay gap against. At present, the verdict is shown only when the two match; otherwise the pay gap displays without a significance mark. This capability is currently available to a limited number of customers, with broader availability planned. |
Release v2026-09-09_01
| Type | Jira | Description |
|---|---|---|
| PA-8378 | On the Pay Equity Overview Report, the Excel, PowerPoint, print, and delete buttons were moved behind a single actions menu, narrowing the action row so it overflows less often in languages with long labels such as German. Deleting a report is now confirmed in a centered dialog instead of a small pop-up, and a failed export shows an error message rather than no response at all. | |
| PA-8693 | A Pay Transparency report generated through a bulk export now shows the same report title and worksheet tab name as the same report downloaded individually for that employee, in every language. The exported file's own name is unchanged. | |
| PA-8597 | Viewing a Pay Transparency statement or downloading a completed bulk export is now recorded in the Audit Log, including who performed the action and which employee's data was involved. Bulk export downloads are logged separately from their creation, and the list of disclosed employees can be downloaded from the Audit Log page. |
Release v2026-09-01_01
| Type | Jira | Description |
|---|---|---|
| PA-8573 | Pay Transparency reports generated in bulk can now be produced in a chosen language, defaulting to the language the employee is already working in. Both the file contents and file names reflect the selected language, and the list of generated reports now shows which language each one was produced in. | |
| PA-8625 | Accented and other non-Latin characters, present in 13 of the 26 supported languages, could render in a substitute typeface rather than the application's own font, making affected words appear bolder and misaligned with the surrounding text. All characters across every supported language now render consistently in the application's font. | |
| PA-8517 | In Pay Equity Analysis > Analysis Overview, each headline pay gap figure now carries a label naming the population it covers (Whole population, Whole subgroup, or Visible records only). This way, an employee whose access is restricted to part of the analysis can tell when the adjusted and unadjusted figures describe different groups. Employees with full access to an analysis see no change. | |
| PA-8532 | Employees who have not been granted access to a dataset's demographic variables now see a message explaining they lack sufficient permission, in place of content, when opening that dataset's overview, its analyses, Workforce Analytics, or Value-based Comparison. The Compensation Assistant remains fully usable in these cases, and the dataset and its analyses continue to appear in lists and search as before. | |
| PA-8603 | For employees whose access to the Pay Transparency report was restricted to specific columns, with "grant access to all columns" turned off, the report came back empty even for columns they had been granted. Figures showed as blank or zero, the employee identifier was replaced by a row number, and the bulk export failed for these employees. The report and its bulk export now display the correct figures for every column an employee has been granted. | |
| PA-8624 | The employee list's Excel export can now be downloaded at any size. This includes the largest customer datasets, which previously produced no file at all. Employees are no longer asked to select fewer columns or filter to fewer rows to complete the download, and a visible error now appears if an export fails for any reason. | |
| PA-8633 | The value range comparison view on the Value-based Comparison report was not displaying its comparison tables, particularly on larger datasets. It could also come up blank after logging out and back in or after switching datasets. The view now renders its comparison tables reliably across sessions and dataset changes. | |
| PA-8630 | In Pay Equity Analysis > Analysis Overview, employees whose data access covers only part of the analysis now see the pay gap after raises presented the same way as employees with full access – as the resulting gap, with a note on which demographic group ends up lower – instead of only a change figure with an up or down arrow. The Results box and its Excel download are affected; employees with full access to an analysis see no change. | |
| PA-8651 | In the left navigation, when the menu was too tall for the browser window and a user scrolled down, the white sidebar panel stopped partway down the screen, leaving items past that point drawn directly on the page's gray background (most often the ADMIN group items). The white panel now extends behind the full menu at any window height and scroll position, so every item appears on it. | |
| PA-8650 | In the Compensation Assistant's Suggested compensation breakdown, the colored square marking each component's band sat flush against the component name with no gap, making the label harder to read. A visible gap now separates the square from the name. |