Philippines staffing research
How long do vendor invoice mismatches wait for an owner decision?
Contents
A measurement protocol for invoice-exception queues, separating review time, owner-decision time, and the evidence that must exist before a payment can be released.
Question and scope
This study asks where time accumulates in a vendor invoice mismatch case and whether the case has the evidence needed when it reaches the payment owner. It covers price, quantity, receipt, duplicate, entity, period, tax, and documentation mismatches. The unit is one mismatch case from identification to closure. The question is about queue design and evidence readiness, not about whether the vendor's charge was commercially fair. The scope includes the invoice, the obligation, the receipt or acceptance evidence, the exception classification, the owner decision, and the corrected payment instruction. It excludes invoices that matched on first review and were paid through the normal process. A case that is closed without a ledger change is recorded as unresolved rather than as a fast closure.
Method and evidence frame
Use a retrospective case review with time stamps. Identify the date the mismatch was first recorded, the date the complete packet reached the owner, the date the owner decided, and the date the ledger or payment instruction changed. Report the intervals separately. A single cycle-time number hides the difference between a team that prepares cases quickly and an owner who responds slowly. The evidence frame separates the invoice claim, the internal obligation, the receiving evidence, the exception decision, and the payment action. A complete packet at the owner's desk is the handoff point. A decision without a ledger change is not a closed case. A ledger change without an approval is a control failure, even if the amount is correct.
Population and unit of analysis
The population is all invoice mismatch cases opened during the observation window for the selected vendor population or business unit. The unit is one case. A case with several disputed lines is one unit if it is decided together, but each line is recorded as an attribute so the value and the classification can be analyzed separately. A duplicate invoice that is canceled before payment is still a case. Define whether cases that are reopened after a vendor credit note are included. Include them and link the rework to the original. Exclude cases that were never assigned to the team, but record the count of unassigned cases separately. Report the counts so the reader knows how much of the queue is represented.
Definitions and measures
The primary measure is the median and range of owner-decision time, reported with the number of cases. A secondary measure is the share of cases that reached the owner with a complete evidence packet. A third measure is the share of cases closed with a matching ledger or payment change. Do not rely on the mean alone when a few large or old cases can distort it. Supporting measures include the time to retrieve the obligation, the time to obtain receiving evidence, the number of owner follow-ups, the value at risk, and the share of cases that required a second review. Report counts and denominators for each. A short decision time on an incomplete packet is not a quality signal.
Sampling and collection
Sample across vendors, mismatch types, values, and business units. Include at least one case that waited on receiving evidence, one that waited on a policy decision, and one that was resolved with a credit note. Use consecutive cases where possible and record every exclusion. A sample of only recently closed cases will understate the age of the open queue. Collect the invoice, the obligation, the receiving record, the exception log, the owner correspondence, and the payment or ledger record. Record the observation time for each. If the case was handled in email, preserve the message headers that show the sender, recipient, and time. Do not reconstruct the sequence from a summary written after the fact.
Analysis and reporting
Report the time intervals by mismatch type and value band. A price mismatch may wait on procurement, while a missing receipt may wait on the receiving team, so a combined average can point to the wrong owner. Show the count, the median, the range, and the share of cases that exceeded the team's stated service target, if one exists. Separate observation from interpretation. A long owner-decision time may reflect an absent approver, a genuinely complex case, or a packet that did not contain the needed evidence. The study can show the interval and the packet completeness, but it cannot assign blame without the correspondence and the approval record. Present the evidence and let the owner interpret the cause.
Limits and sources of error
Invoice systems and email archives retain different levels of detail. A case may be discussed in a meeting and recorded only as a status change. The reviewer should record that limitation. Small samples are sensitive to one large vendor or one month-end backlog, so stratify by period and value. Currency differences can make a value comparison misleading if the exchange rate is not stated. The study cannot establish that the invoice was valid, that the payment was lawful, or that the accounting treatment was correct. It measures queue time and evidence readiness. The AICPA, the U.S. Government Accountability Office, and the Internal Revenue Service provide general control and recordkeeping context, not a decision for a specific invoice.
Operational use
Use the result to separate preparation from approval, add a completeness check before the owner review, or set a follow-up rule for cases that wait too long. A bottleneck concentrated in one owner or one evidence source is a process design issue. Fix the handoff rather than asking the administrator to work faster on an incomplete packet. A clear time measure also helps the vendor conversation. The team can say what evidence is missing and when the next update will come, instead of sending a generic message that the invoice is under review. That reduces repeated follow-ups and makes the queue more predictable.
Review cadence
Review the mismatch queue monthly and repeat the measurement after a system, vendor, or approval-policy change. Keep the interval definitions stable so the comparison is meaningful. If a new approval layer is added, label the period and do not compare the new cycle time directly with the old one. Close the review with a named owner, one change, and a re-test date. Preserve the sampled cases and the time stamps so a later reviewer can repeat the measurement. A queue that looks fast because cases are closed without a ledger change is not a controlled queue.