Methodology
This study reconstructs the AP close as a sequence of dated events rather than a checklist of tasks. The question is where time is spent, which dependencies control the sequence, and whether the evidence at each handoff supports the next step. The unit is one close cycle for one entity, with each invoice or accrual item available as a sub-unit when the cycle needs to be explained. The method does not set a benchmark number of days, and it does not treat a late posting as proof of poor performance without looking at the dependency that caused it.
Evidence and scope
The population is the close activity for a declared entity and period. Sources include the close checklist, ERP posting timestamps, the receipt and invoice registers, the accrual working file, the payment run log, the reconciliation sign-off, and the reversal record for the following period. Record the extract time and reconcile the checklist item count to the workpapers actually produced. A task marked complete with no artifact is a finding. A task performed but not marked complete is also a finding, because the timeline will otherwise show a delay that did not happen. Record who supplied each manual timestamp and whether it came from a person or a system export.
Adverse controls
Several conditions distort a timeline. Timestamps may come from systems in different time zones. A checklist may be updated at the end of the close rather than as work happens. An approval may sit in email and be entered later. Two people may work in parallel on the same reconciliation without a shared version. A late invoice may be backdated to the prior period by a manual entry. The method compares system timestamps with the business date and records the source of each event. Where only a manual time is available, the event is marked as reported rather than observed.
Analysis
Analysis divides the cycle into segments: invoice cutoff to population freeze, freeze to accrual proposal, proposal to controller review, review to posting, posting to reconciliation sign-off, and sign-off to the next-period reversal. For each segment, report the elapsed time, the owner, the dependency, and whether the work was waiting or active. Then identify the critical path, meaning the chain that determined the final posting date. Report the median and the tail, and show the items that moved between segments more than once. A single average close duration hides the rework that makes a cycle unpredictable.
Research-to-practice
For practice, the timeline should be a working control, not a retrospective chart. Give each event a named owner, a required input, an output artifact, and an acceptance test. The population freeze should have a row count and a control total. The accrual proposal should have a method note and a controller review. The reconciliation should show open items with owners. The reversal plan should name the next-period comparison. When an event is late, the team should be able to see which upstream input was missing rather than recording a general delay. A timeline that is not used during the close tends to become a retrospective narrative, and the dependencies are reconstructed after the fact.
Independent validation
Independent validation reconstructs one completed cycle from the artifacts, without relying on the checklist narrative. The reviewer selects a documented sample of items, traces each from invoice or receipt to posting and reversal, and compares the reconstructed timestamps to the reported timeline. Differences are logged by event and by source. If the reviewer cannot reproduce the sequence, the timeline is a summary rather than evidence. The review should include at least one manual journal entry and one accrual reversal, because those are the points where the system record is most likely to be incomplete.
Decision-log design
The decision log should carry a cycle identifier, entity, period, event name, planned date, observed timestamp, source system, owner, input artifact, output artifact, acceptance result, dependency, and delay reason. For item-level tracing, add invoice or accrual identifier, amount, currency, posting reference, reversal reference, and reviewer. A delay reason should be coded, such as missing receipt, late approval, system outage, unresolved dispute, or reviewer availability. Free text can explain the code but should not replace it, or the report cannot be aggregated.
Buyer review checks
Buyer review checks should ask how the provider defines the start and end of the close, which timestamps are system-observed, and how it handles work performed outside the checklist. Ask to see a cycle where an item slipped and the evidence that explains the slip. Ask how the provider separates waiting time from active work and how it protects the controller review from becoming a rubber stamp. A provider that reports only that the close was completed on time is not showing the dependencies that make the next close predictable.
Field dictionary
Field definitions should distinguish planned from observed time. Planned date is the close calendar target. Observed timestamp is the system record of the event. Reported time is a manual entry without a system trace. Population freeze is the point after which new invoices do not enter the current accrual, even if they are later posted to the period under an approved rule. Accrual proposal is the line-level estimate sent for controller review. Posting is the ledger entry. Sign-off is the reviewer's acceptance. Reversal is the next-period entry that clears the accrual. Each term should map to a specific artifact.
Sampling plan
The sampling plan selects one or more complete cycles plus boundary items: the largest accrual, the oldest open item, a manual journal entry, an item posted after the freeze, a reversal that did not match its accrual, and an item that moved between owners. Select the items before reviewing outcomes, and state the number selected. The plan tests whether the timeline can be reconstructed from evidence; it is not a statistical estimate of close performance across the company. Include a cycle with a manual journal entry, because manual entries often sit outside the automated timestamps.
Release and rollback gate
Release and rollback apply to a new close timeline or tracking tool. Pilot it on one entity while the existing checklist remains in place. Acceptance criteria belong to the controller: every event has an observed or reported source, the population reconciles, the accrual proposal has a method note, and the reversal can be matched to the accrual. If the tool cannot capture the manual steps, roll back to the prior checklist and record the gap rather than presenting a partial timeline as complete.
Training and continuity
Training should cover event definitions, the difference between observed and reported time, the population freeze, and the reversal comparison. Use a worked cycle with a late receipt, a manual entry, and a disputed invoice. Ask participants to identify the critical path and the evidence they would need to prove it. Recalibrate after an ERP upgrade or a change in the close calendar, because those changes alter both the timestamps and the dependencies. A short event dictionary is more useful than a long narrative procedure. Ask each participant to place the events on the timeline independently, then compare the versions before discussing the cause.
Limitations
Limitations include incomplete system timestamps, parallel work that cannot be ordered from the records, and manual entries that are documented after the fact. The method observes one entity and period and cannot establish a company-wide norm. Seasonality, acquisition activity, and system outages can change the critical path without changing the process. Public sources on records management and internal control support the value of dated evidence and review, but they do not prescribe a close duration. Any target should come from the company's own obligations and risk tolerance. The method should also state whether the timeline covers one legal entity or a consolidated view.
Conclusion and refresh trigger
The conclusion is that an AP close timeline is useful when it explains dependencies, not when it assigns blame. A cycle is reviewable when each event has an owner, an input, an output, an acceptance test, and a timestamp whose source is known. Refresh the timeline when the ERP, close calendar, approval workflow, or payment run schedule changes, and when the same segment repeatedly determines the critical path. The refresh should begin with the event dictionary and the population freeze, because those two choices decide whether the rest of the timeline can be compared across periods.
Methodology and limitations
This report synthesizes the cited primary and institutional sources, then applies their published control principles to the bounded accounts payable question described above. The evidence is contextual. It does not measure this company, guarantee an outcome, or replace legal, security, accounting, or payment judgment.
Map the AP close timeline
Bring the close checklist, system timestamps, accrual file, and reversal record. The session identifies the critical path and the evidence gaps.
Plan the close reviewSources
These primary sources support the control principles and evidence boundaries in this report.
FAQs
Are the planning examples benchmarks?
No. They describe a testable workflow shape and are not promises, market averages, or production targets.
What should an outsourced AP specialist 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.