What does data coherence check?
Data coherence flags certain unusual values in your HR and recruitment data. It helps you find the employees or applications concerned and check their data in your source software.
💡 This page is available to Reflect administrators. The checks shown depend on the modules enabled and the data available in your account.
An alert calls for a check
A flagged value can be correct: for example, a 60% base salary increase after a change of role. Conversely, a data entry error can remain below the thresholds and trigger no alert. These checks therefore do not detect every payroll or data entry error.
Indicators and detection thresholds
The page can show three checks, depending on your modules and the data available. A value that crosses the threshold is flagged in red.
Field checked | Value flagged |
Base salary increase rate | Strictly greater than 50%. |
Time to hire (recruitment lead time, in days) | Strictly greater than 100 days. |
Arrival date | Strictly before 31 December 1950. The label displayed is “Before 1950”. |
A 50% increase or a Time to hire of 100 days does not trigger these alerts. A green result only means that no value crosses the threshold for the selected period and filters.
Find and investigate flagged values
Open Data management → Data coherence, or go directly to the Data coherence page.
Click the red check row to see the people concerned, the flagged value and the affected metrics. The Time to hire check concerns applications; the other two concern employees.
If the list is longer than the preview, click Open in explorer to view the full list. Check the period and filters, then add any columns you need for your investigation.
Export the list to Excel if you need to send it to the person responsible for correcting the data in the source software.
How do you correct an inconsistency?
Compare the flagged value with the employee or candidate record in your HRIS, payroll software or recruitment system.
If it is wrong, correct it in the source software. For a file-based source, also correct the file sent to Reflect.
Wait for the next update of that source, or submit the corrected file using your import method. See the Data update guide.
Return to the same check, with the same period and filters, to verify the result after the update.
💡 If an exceptional value is correct, keep it. The alert remains visible as long as the value crosses the threshold; there is no need to change correct data just to remove an alert.
Filter to focus your checks
The + button at the top of the page lets you filter data coherence checks, for example by entity or department, depending on the available filters. If Incompatible selected filters appears, remove the filters that do not apply to that check.
Choose the period to check
The period in the top right determines which salary increases, hires and arrivals are included in the checks. Expand it to look for older cases.
Changing the period or filters focuses your check: it does not correct or delete source data, or cancel its effect on your other analyses.