On this page Jump to section
Abstract. This paper describes a small study for measuring how long it takes to retrieve the supporting document behind a QuickBooks transaction. It covers sampling, timing boundaries, failure states, and the reporting format that makes the result useful to a bookkeeping team.
The study is descriptive. It does not claim that a particular retrieval time is good or that a tool will produce a specific improvement.
The question behind the delay
Month-end close is usually discussed as a posting problem. In practice, a large share of the delay is documentary: the transaction exists, but the invoice, receipt, or statement behind it is missing or hard to find.
A retrieval-time study asks: given a transaction in the close population, how long does a competent reviewer need to produce the supporting document, and how often do they fail?
That question produces a number the team can act on. It also shows where the archive is weak.
Define the population
Use the transactions sampled for the close review, not the whole ledger. The close population is the set of items a reviewer would actually check.
Record the period, the accounts included, and the sampling rule. Include bills, expenses, deposits, and journal entry support in proportion to their review weight.
Exclude transactions with no expectation of support, such as bank transfers or tax payments, unless the close policy says otherwise.
Set the timing boundaries
Start the clock when the reviewer receives the transaction identifier and the request. Stop when the document is open and verified against the amount and date.
Include the time spent searching the accounting system, the shared drive, email, and vendor portals. Do not include time spent reading the document for content review, which is a different task.
Record failed searches separately. A search that ends in "not found" is a result, not a missing observation.
Sample and stratify
Draw a sample large enough to describe the process, not the population to three decimal places. Twenty to forty items is workable for a small team.
Stratify by document type and source. An emailed receipt and a vendor portal statement have different retrieval paths, and averaging them hides the problem.
If the team already knows one source is weak, oversample it. The study should inform a fix, not just produce a balanced average.
Control the conditions
Use the same reviewer, or two reviewers on a shared subset, and the same environment. A study run during a busy close will measure interruption, not retrieval.
Note whether the reviewer knew the transaction in advance. Familiarity shortens search time and flatters the result.
Avoid timing the same document twice in one session. Memory turns a search into a recall test.
Record the failure modes
For each failed or slow search, record the cause: no attachment, wrong attachment, unreadable file, wrong naming, missing access, or document never received.
These categories drive different fixes. A naming problem is training. A missing-access problem is permissions. A document-never-received problem is a vendor or intake issue.
Report the distribution
Report the median, the 75th and 90th percentiles, and the failure share. Averages hide the long searches that actually delay a close.
Add the estimated total hours for the close by multiplying the median by the review count, and show the range. State the assumptions clearly.
Use the result
If retrieval dominates, fix the archive before adding review capacity. A naming standard and an intake checklist often reduce the median more than another reviewer.
If failures cluster around one vendor, change the intake path for that vendor. If they cluster around one account, check whether the documents are being filed at all.
Re-run the study after the change with the same protocol so the comparison is valid.
Limitations
The study measures one team in one period. It is not a benchmark.
Timing introduces observer effects, especially when the reviewer knows the purpose. Triangulate with attachment counts and failed-search logs where available.
The protocol also assumes the document exists somewhere. If the business never collected it, retrieval time is not the right metric; intake completeness is.
Sources
- PCAOB Auditing Standard 1105, Audit Evidence: https://pcaobus.org/oversight/standards/auditing-standards/details/AS1105
- Committee of Sponsoring Organizations of the Treadway Commission, Internal Control over Financial Reporting guidance: https://www.coso.org/guidance-on-ic
- Internal Revenue Service, How long should I keep records: https://www.irs.gov/businesses/small-businesses-self-employed/how-long-should-i-keep-records
- U.S. Government Accountability Office, Standards for Internal Control in the Federal Government: https://www.gao.gov/green-book