Check the files
PaidWhen checks required columns, dates, positive amounts, duplicate invoice IDs, and whether there are enough usable closed invoices.
PaidWhen checks whether your own payment history contains a repeatable pattern before asking your team to trust the ranked list.
The payment date is used to check what happened. It is never used to predict an open invoice.
PaidWhen checks required columns, dates, positive amounts, duplicate invoice IDs, and whether there are enough usable closed invoices.
A closed invoice is marked late when its payment date comes after its due date.
For each historical invoice, PaidWhen uses only account behavior that was already known at that time.
Older invoices are used to learn the pattern. The newest quarter is kept aside to see how that pattern performs on later business.
The result must be more useful than assigning every invoice your company’s usual late-payment rate.
Each open amount is multiplied by its estimated chance of delay. The resulting cash-at-risk value sets the order of the list.
The workspace includes plain explanations beside the technical measurements.
The estimated chance that an invoice will be paid after its due date.
Invoice amount × late risk. It combines likelihood with financial impact.
How well the check put later-paid invoices above on-time invoices.
How close the estimated percentages were to what actually happened. Lower is better.
A run is marked limited or unavailable when there are too few invoices, too few examples of both outcomes, too little recent history to test, or no improvement over the simple comparison.
The sample workspace shows the list and its reliability measures.