Research · Published:
Which purchase order version should AP use? A chronology study
Research on purchase order amendments, effective timing, invoice comparisons, and decision ownership in outsourced accounts payable.
Research question and method
Research question: which purchase order version was authoritative when an invoice entered review, and can another person reconstruct that answer from retained records? This study examines amendments because a current purchase order can conceal the terms visible to accounts payable at an earlier cutoff. The method maps GAO principles for quality information and documented control activity, NIST controls for access and audit records, and National Archives records-management guidance to a bounded AP sample. Cases include a price change, quantity increase, closed line reopened, supplier substitution, amendment after invoice receipt, and an urgent request supported only by email. The study does not decide contractual rights. It tests whether source versions, effective times, and employee decisions remain visible while an outsourced AP specialist prepares a comparison.
Evidence population
The evidence population begins with the original purchase order and every retained amendment. For each case, record entity, supplier identifier, order and line, version identifier, changed fields, requester, authorized approver, effective time, system event, invoice receipt time, receipt evidence, and exception disposition. Values drawn directly from a source are facts. Saying the invoice amount differs from version three is an analysis when both values and the comparison rule are stated. Deciding that version three governs the invoice is a company decision unless an approved rule resolves the case without interpretation. The preparer should preserve that distinction in the packet. A current screen capture alone is insufficient when the system overwrites former values or does not show when an amendment became effective.
Case analysis
Chronology is tested at uncomfortable boundaries. One invoice may arrive before an amendment but reach approval afterward. Another may cite a revised unit price that purchasing approved orally before the system changed. A third may match the latest order while exceeding the quantity received at the review cutoff. The packet should display each event on one timeline without treating later evidence as if it existed earlier. An email can explain why a change was requested, but it should not silently substitute for the designated approval record. If company policy allows retrospective amendments, the responsible employee should record that decision and its authority. Outsourced support may link the events and calculate differences. It should not backdate a change, choose an effective version, or clear the resulting exception.
Reproduction test
The reproduction test gives two reviewers the same version history and invoice packet. Each reviewer identifies the order values visible at invoice receipt, first preparation, approval, and any later correction. They must cite the source event for every value and flag missing or inaccessible history. Compare their results for reopened lines, cancelled amendments, time-zone differences, and changes approved in one system but synchronized later to another. A disagreement is a finding about evidence or policy clarity. It should remain in the study record with both interpretations rather than being resolved by selecting the cleaner timeline. Repeating the exercise after an amendment tests whether the workflow preserves the former state or merely presents the newest record as timeless truth.
Operating boundary
Access and authority require separate treatment. An outsourced Philippines-based AP specialist may retrieve approved order records, compare invoice fields, maintain the exception timeline, request missing evidence, and route a factual note. Authorized company employees retain purchasing changes, receipt acceptance, accounting treatment, invoice approval, exception override, and payment release. NIST least-privilege principles support limiting the preparer to the records and functions needed for that lane. Technical permission to edit an order does not create business authority. If the system cannot give read-only historical access, the organization should document the compensating review rather than expecting support staff to avoid an available function by memory alone.
Interpretation
Useful outcome labels describe evidence, not payment decisions: version history complete; amendment awaiting authorized approval; effective-time conflict; invoice predates amendment; synchronization delay under review; source history inaccessible; or employee disposition recorded. The queue should name the present owner and last evidence event. A closed ticket count cannot show whether the correct version was used, and a long-open item may reflect a deliberate stop. This research does not propose a time target, variance tolerance, or automatic precedence rule. It proposes fields that allow a finance owner to apply company policy consistently and allow a later reviewer to see where judgment entered the process.
Limitations
Limitations: the cited public sources describe internal control, system access, and recordkeeping at a general level. They do not interpret a private purchase agreement, establish when an amendment becomes legally effective, set accounting treatment, or identify a valid approver for a particular company. Systems record versions and timestamps differently, and migrations may lose history. A selected exception sample cannot estimate how often incorrect versions affect invoices. The test may also miss oral decisions or records held outside approved systems. Those gaps must be listed rather than converted into assumptions. Management, procurement, accounting, and counsel remain responsible for their own policies and obligations.
Evidence-led conclusion
Evidence-led conclusion: AP can prepare a defensible order comparison when the record retains each version, its approval and effective time, the invoice cutoff, and the employee disposition. The current purchase order is not enough if it erases what reviewers could see earlier. Outsourced support adds value by assembling the chronology, showing field-level differences, and keeping unresolved questions assigned. It must stop before deciding which amendment governs or altering the purchasing record. A successful reproduction test lets a new reviewer follow the same events and identify the same unresolved point without oral explanation. That is the useful control result, even when the commercial decision remains open.
Sources
These primary sources support the control principles and evidence boundaries in this report.
FAQs
Are the planning numbers benchmarks?
No. They describe a testable workflow shape and are not promises, market averages, or production targets.
What should an outsourced AP assistant own?
Repeatable preparation, documentation, status tracking, and follow-up within least-privilege access. Named finance owners retain approval and payment decisions.
When should an item be escalated?
When evidence is missing, a request changes payment details, a duplicate or fraud signal appears, or the item falls outside the written rule.