Blog · Procurement and supply chain · Procurement
Identify dated purchase-order sequence exceptions while separating timestamp problems, approved exceptions and unresolved policy evidence.
Review after-the-fact purchase orders by their recorded event sequence. Compare a known approval event with the linked invoice and receipt evidence, then ask why the order was approved later. Chronology identifies a case to review; it does not establish that a requester broke a rule or that the supplier should not be paid.
Microsoft's requisition workflow documentation distinguishes requisition approval from subsequent order generation. That is one example of why a request approval, PO creation and PO approval must not be silently treated as the same timestamp.
Use one reviewed PO version or line as the approval record, linked to its invoice and receipt evidence. Choose the grain explicitly: an order amended after an initial valid approval can otherwise appear entirely retrospective when only one revised line is in question.
Inputs include company, order key and version, approval timestamp, creation timestamp, invoice key and document date, invoice receipt timestamp if available, product or service acceptance event, exception policy, reviewer evidence and cutoff time zone. Retain the original date fields and their documented meanings.
Start with the PO, invoice and paid-spend population rules. Invoice document date, AP entry date and payment date describe different events and cannot be interchanged to make a sequence look complete.
Possible tests include PO approval later than the invoice document date, later than recorded invoice arrival, or later than the linked receipt event. Report each separately. A supplier-issued invoice date is not always the moment the organization received it.
When timestamps are precise and comparable, calculate the elapsed difference. With dates only, report calendar-day differences and identify same-day order as unresolved if time order matters. Do not manufacture midnight approval times to create precision the export lacks.
The applicable purchasing policy determines eligibility and permitted exceptions. Some recorded events may use a different approved process. Keep that documented status separate from the mechanical chronology flag.
These records and USD invoice amounts are synthetic. All dates occur in October 2026. Approval evidence is date-only; the report cutoff is October 6. Calendar-day lags compare the displayed dates, not working days.
| DEMO case | Receipt date | Invoice date | PO approval date | Invoice amount | Initial chronology |
|---|---|---|---|---|---|
| A | Oct. 5 | Oct. 4 | Oct. 3 | $2,000 | Approval before both events |
| B | Oct. 2 | Oct. 3 | Oct. 5 | $3,000 | Approval 3 days after receipt, 2 after invoice |
| C | Oct. 4 | Oct. 1 | Oct. 6 | $1,500 | Approval 5 days after invoice |
| D | Oct. 5 | Unknown | Oct. 4 | $2,500 | Invoice sequence unknown |
The invoice population totals $9,000. B and C form two invoice-date chronology flags totaling $4,500; A is known earlier and D cannot be evaluated for that test. The $4,500 is flagged invoice value, not a confirmed policy violation or financial loss.
Suppose C has reviewed evidence of an applicable approved exception. Retain C's late sequence while classifying its policy disposition separately. The decision is to investigate B's process history, retain C's exception evidence and obtain D's invoice date; erasing C would make the chronology report less faithful.
Check migrated orders, imported historical approvals, corrected document dates, canceled orders and replacements. A system migration timestamp can describe data entry rather than an original authorization event. Record that distinction instead of classifying every migrated record as late.
Avoid repeating the full order value for each invoice. Several invoices may link to one PO, and each may have a different sequence. Report the affected line or invoice value on a consistent basis and reconcile it to the source population.
The maverick spend review concerns buying outside the reviewed channel. A late order can be one signal but does not prove that broader classification. Check whether the issue is date evidence, an approved exception, requisition routing or genuinely retrospective authorization.
Requisition approval delay analysis can help examine upstream workflow timing. Keep the chronology review narrow enough that it identifies a concrete process question rather than assigning unsupported intent to staff.
Show known-evaluable records, flagged records, approved exception dispositions and unknown evidence. Give each unresolved case a responsible reviewer and evidence request. Preserve earlier snapshots if later approvals or corrections change the result.
The procurement review example illustrates a bounded report with visible limitations. Contact Covirage to discuss a sequence review using available source dates and the organization's approved policy definitions.
No. It flags chronology under the chosen rule. Review date meaning, approved exceptions and the applicable purchasing policy before classifying it.
Only if that equivalence is documented in the source process. Otherwise creation and approval are separate milestones, and absent approval evidence remains unknown.