The Past Date Threshold analysis scans date and date-time fields, and flags any value that exceeds a configurable age limit such as "more than 5 years in the past." The cutoff date can be a fixed date or relative date. 

Use this analysis to surface stale records, catch data entry errors, and identify objects where lifecycle processes may have broken down.

This analysis marks past dates (beyond a certain threshold date) as an anomaly by default. If you have date fields where past dates are the norm, and you want to scan for dates that are incorrectly in the future, you can either invert the configured threshold for those fields, or use the Future Date or Date Range Analysis instead. 
In a future version it will be possible to get notified at certain signal thresholds, opening up more use cases for this analysis type.

Configuration

Set a date as a threshold. The analysis will search for date values that are older than that date, so the date itself is not included.

At Date Input Type, select ‘Fixed Date’ for a date that stays the same, or ‘Date Literal’ for a relative date that is recalculated with each analysis run. 

At Threshold Date, 

• For Fixed Date, type a date or select it with the Calendar button. 
• For Date Literal, enter a relative date. For example, use LAST_N_DAYS:100 to find date values dating back to more than a hundred days ago. Or use LAST_YEAR to find date values dating back to earlier than last year: if today is 5 June 2026, and you use LAST_YEAR, the threshold date will be 1 January 2025, and the analysis results will include record values of 31 December 2024 and earlier. 
  Each time you use the analysis with this date literal, the threshold date is recalculated. 
  Read more about how to format Date Literals here.

At Thresholds, set signal thresholds for what constitutes a good, warning level, or critical past date rate. It depends on your use case and the type of field how good or bad it is to have fields dating back in the past. 
If, for a certain field, you want to have as few records beyond a certain past date as possible, you could for example set a good past date rate at 5% or less, warning level would be between 5-20%, and anything over 20% would be critical.
For those cases where having records beyond a certaina date in the past is regarded as good, you can reverse the threshold signalling by clicking the Reverse button. 

Detailed Job Results

The bar chart displays the percentage of records that have a date beyond the set date threshold, per field. Hover over a bar to see the number of records, and the percentage of all records with a value in that field.

The doughnut chart shows which share of the found records has a date value close to the threshold date (e.g. no more than a year older than that date, or ten years or more) or in an earlier time range. The time ranges are calculated from the threshold date.

The results table shows the total number of records containing a value in each analyzed field, the number of records with a date beyond the set date threshold, the percentage of records with a date beyond the set date threshold out of all records containing a value, and the date threshold used in the analysis. The percentage of records with a date beyond the set date threshold is color-coded by the signal threshold settings (green for Good, orange for Warning, red for Critical).

Key Insights

• Data plausibility: Identify which date fields contain unrealistically old values that are unlikely to reflect real business activity.
• Stale or unupdated records: Surface records that should have been closed, updated, or archived but were not, such as open Opportunities or unresolved Cases with very old dates.
• Migration and import issues: Highlight records where historical dates from a migration or data import may be incorrect or out of range.
• Process and integration failures: Detect patterns where users or integrations are (intentionally or unintentionally) backdating records beyond what is expected, or where lifecycle transitions (stage changes, closures) are missing.

Recommended Actions