Month End Close Checklist: Nine Steps and Where the Days Go

A month-end close checklist turns a month of transactions into statements someone can sign: cut off the period, reconcile, accrue, review and lock. The close takes as long as its longest chain of waits, not its longest list of tasks. That is why automation can save hours but no days, and why a faster close estimates more of the month.

The month-end close checklist runs in nine steps, from cut-off to the true-up log

This month-end close checklist is the minimum a startup needs before its numbers can feed the monthly comparison with budget or a board package. Its 9 steps run in the order their inputs arrive.

1. Cut-off. Before month end, tell every department the day accounting stops booking actuals and starts accruing.
2. Cash. Run the bank and card reconciliations, payment processor included, against statements or complete feeds.
3. Revenue. Post every customer invoice for the month, then turn billings into revenue and deferred revenue on the revenue recognition schedule that ASC 606 requires.
4. Payroll. Book wages, employer taxes and benefits for days worked; accrue unpaid bonuses and commissions.
5. Accruals and prepaids. Accrue costs incurred but not yet invoiced, from contracts and run rates; spread prepaid costs; book depreciation.
6. Balance sheet. Tie receivables, payables, accruals, prepaids and deferred revenue to supporting schedules; read the ageing reports.
7. Review. A second person explains every movement above a set threshold against last month and budget.
8. Sign-off. Approve the package, lock the period and file the workpapers.
9. After the close. Log each true-up and every entry later posted to the closed month.

APQC’s survey of 2,300 organisations put the median monthly close at 6.4 calendar days, with the slowest quarter needing 10 or more (Wiggins, CFO.com, 2018). Two companies with the same checklist can sit at either end of that range. Exhibit 1 shows why: beside each step it lists what the step must wait for.

Exhibit 1. The month-end close checklist: what each step proves and what it waits for

Step	What it proves (audit assertion)	Waits for	Ten-day close	Three-day close
1. Cut-off	Cutoff	A cut-off notice sent before month end	Day 0	Day 0
2. Cash	Existence, completeness	Statements or complete bank feeds	Days 1 to 4	Daily in the month; last day on day 1
3. Revenue	Occurrence, cutoff, accuracy	Billing export, contract changes	Days 1 to 2	Day 1, from the billing system
4. Payroll	Completeness, accuracy	The payroll provider’s report	Days 1 to 2	Day 1, accrued from the pay calendar
5. Accruals and prepaids	Completeness, cutoff	Vendor invoices	Days 6 to 7, from invoices	Day 1, from contracts and run rates
6. Balance sheet	Existence, completeness, valuation	Steps 2 to 5	Day 8	Day 2
7. Review	All of the above	The reviewer’s calendar	Day 9	Day 3, by exception
8. Sign-off		Steps 6 and 7	Day 10	Day 3
9. After the close	Cutoff	Invoices that arrive later	Next month	Next month

Source: assertions from ISA 315 (Revised 2019), paragraph A190. Timings are the hypothetical company of Exhibit 3, in business days after month end.

A month-end close lasts as long as its longest chain of waits

Project managers have had a name for this since the late 1950s, when Morgan Walker of DuPont and James Kelley of Remington Rand developed the critical path method (Critical path method, Wikipedia; Kelley, 1961). A project is a network of tasks, each with a duration and a list of tasks it depends on. Its length is the longest chain of dependent tasks. Every task off that chain has float: it can slip, or finish early, without moving the end date. The method also measures drag, which it defines as “the amount by which the project’s duration is extended by each critical path activity and constraint”.

A constraint can be a wait rather than work. The textbook case is a solar-panel test that “would not start until the scheduled time for sunrise”. A month-end close is full of sunrises: the bank statement, the payroll provider’s report, the invoice a vendor sends on the sixth, the reviewer who is free on Thursday.

The benchmarks count those waits. APQC defines its close cycle time as “the total time from the beginning of the process to the end, including time spent actually performing the process as well as time spent waiting to move forward”. In Ledge’s 2025 survey of 100 finance professionals, the obstacle cited most was “Dependency on other departments and regions”, named by 56%, ahead of transaction complexity at 39% (Ledge, 2025). Ledge sells close software and does not say how it chose respondents. Ashutosh Deshmukh and Maximilian Holzmeier, surveying 55 German corporations on their annual close, list among the top practices: “Move routine work or noncritical activities out of the closing time period” (Deshmukh & Holzmeier, Strategic Finance, 2023).

Exhibit 2. Close benchmarks measure different clocks

Source	Sample and year	Clock	Finding
APQC (non-profit benchmarking body)	2,300 organisations; published 2018	Calendar days, trial balance to consolidated statements	Median 6.4; top quartile 4.8 or less; bottom quartile 10 or more
Ventana Research (analyst firm)	Sample not stated; 2019	Business days, monthly close	61% within 6 (53% in 2015); 46% within 4 (29% in 2015)
Ledge (close-software vendor)	100 finance professionals at companies of 51 to 10,000+ employees; 2025	Business days	18% in 3 days or less; half take longer than 5
Deshmukh & Holzmeier	55 German corporations; 2019	Days, annual close	Most take 11 to 30; 84% call their close length appropriate

Source: as listed; all figures are self-reported. Ramp’s month-end close guide attributes the 6.4-day median to a PwC report; the figure is APQC’s.

Automating a step off the critical path saves hours, not days

The company in Exhibit 3 is hypothetical: a seed-stage SaaS business with one bank, two corporate cards, a payroll provider and a part-time controller, closing in ten business days. The controller waits six days for vendor invoices before booking accruals, and a day for statements before three days of cash reconciliation; payroll, the balance sheet, the review and the package take a day each.

Exhibit 3. Shortening a hypothetical ten-day close, one change at a time

Change	Close length (business days)	Days saved	Critical path afterwards
Baseline	10		Vendor invoices (6-day wait), accruals, balance sheet, review, package
1. Automate bank and card matching (3 days to half a day)	10	0	Unchanged: cash reconciliation had float
2. Accrue vendor costs on day 1 from contracts and run rates; true up when invoices arrive	5	5	Payroll report (1-day wait), payroll entries, balance sheet, review, package
3. Accrue payroll from the pay calendar	4.5	0.5	Bank statements (1-day wait), cash matching, balance sheet, review, package
4. Reconcile cash daily during the month	4	0.5	Accruals and payroll entries (tied), balance sheet, review, package
5. Review by exception, with the reviewer booked in advance	3	1	Accruals, balance sheet, then review and package on day 3

Source: our critical-path calculation for a hypothetical company; durations and dependencies are assumptions.

The first change is the one most close software is sold on, and here it saves nothing. Cash reconciliation finished on day 4 and its output was not needed until day 7, so it had 3 days of float. The change that matters removes a wait: the controller stops waiting for invoices and estimates what has not yet been billed. After that the critical path moves twice in two steps, first to the payroll report and then to the bank statements. Shorten the longest chain and another one becomes the longest.

None of this argues against automation; it sets the order. Ventana’s 2014 research found that 16% of companies making substantial use of spreadsheets closed within four business days, against 32% of those that limit them (Ventana Research, 2014). Its 2019 data put 88% of heavily automated companies within six business days, against 40% of those with little or no automation (Kugel, 2019). Both are correlations from samples of unstated size, and the first was paid for by a close-software vendor. Automating in the wrong order buys the software and keeps the ten-day close.

Cut-off and accruals turn waiting into estimating

But a controller who stops waiting for invoices has to put a number on costs nobody has billed yet. One of the assertions auditors test, in the international standard on risk assessment, is cut-off: transactions and events “have been recorded in the correct accounting period” (ISA 315 (Revised 2019)). APQC’s advice on cut-off makes the switch explicit: tell every department “when accounting will stop booking actuals and accruals will start being recorded”. A cut-off is therefore a decision about when estimating begins.

Payroll shows why this matters for a startup. Kruze Consulting’s client data put payroll-related costs at 76% of venture-backed startups’ operating expenses (Kruze, 2024), and pay periods rarely match the calendar month. In BLS data for February 2023, 43.0% of US private establishments paid biweekly and 27.0% weekly, so 70.0% ran pay periods that cannot end on the last day of every month (our arithmetic). The figures count establishments, not employees, from a survey of about 122,000 businesses and government agencies (BLS, 2023).

Accruing from the pay calendar books payroll on day 1, a day or two before the provider’s report arrives, at the cost of a small tax true-up later.

A three-day close reports younger numbers, and more of them are estimates

A month averages 30.4 days, so the average transaction in a monthly package is about 15.2 days old at month end. A package delivered on day 10 describes a month whose average day lies 25.2 days back. On day 3 the figure is 18.2 days, with 27.4 days of the month left to act on it instead of 20.4. No rule demands this. Large accelerated and accelerated filers have 40 days to file a quarterly report, and other listed companies 45 (SEC Form 10-Q). An optional clause in the National Venture Capital Association’s model investors’ rights agreement gives a startup 30 days for monthly statements (NVCA, 2023).

The cost rises as the target falls. In 2001 a PricewaterhouseCoopers partner, Steve Lukens, told CFO magazine: “You can get to five days without having to make huge sacrifices, but to get from five days to two or one is much more difficult” (CFO.com, 2001). In Exhibit 3, five and a half of the seven days saved came from estimating sooner. Patricia Dechow and Ilia Dichev note that “accruals require assumptions and estimates of future cash flows”, and argue that “the quality of accruals and earnings is decreasing in the magnitude of estimation error in accruals” (Dechow & Dichev, 2002).

Cisco shows the limit of speed. Its chief executive, John Chambers, defined the virtual close in 1999 as “the ability to close the financial books with a one-hour notice”. On 8 May, after a fiscal quarter that ended on 28 April, the company reported quarterly net sales of $4.73 billion and a net loss of $2.69 billion (Cisco, 2001). Its annual report named the largest item, inventory bought against forecasts of demand that then fell: “The excess inventory charge recorded in the third quarter of fiscal 2001 was $2.25 billion”, equal to 48% of the quarter’s net sales (Cisco 10-K, 2001). A fast close reported the loss quickly; it did not make the forecast behind it any better.

A fast close can also be fast only on paper: Robert Kugel of Ventana Research reported anecdotal evidence in 2019 of companies posting material corrections in the weeks after their reported close (Kugel, 2019). We call it the last-entry day: the latest date on which any entry dated in the closed month was posted. A company that sends its package on day 3 and keeps posting to that month until day 12 runs a 12-day close.

Its companion is a true-up log, which records each accrual, the invoice that later settled it, and the difference. A company that accrues $80k of vendor costs on day 1 and later receives invoices for $86k carries a $6k difference, 7.5% of the estimate, into the next month’s results. A year of these entries shows whether the estimates are good enough to close on day 3.

Review and sign-off are the steps to keep in human hands

Faster estimates put more weight on the step built to catch bad ones. Listed companies make the review personal: under section 302 of the Sarbanes-Oxley Act, the chief executive and the chief financial officer each certify that “the signing officer has reviewed the report” (15 U.S.C. § 7241). Weili Ge and Sarah McVay studied 261 listed US companies that disclosed material weaknesses after Sarbanes-Oxley. The weaknesses tended to involve revenue recognition, segregation of duties, the period-end reporting process and account reconciliation (Ge & McVay, 2005). Fastly’s 2019 prospectus disclosed such weaknesses for 2017 and 2018, tied to the “lack of sufficient qualified accounting personnel” and to its close controls. Part of the remedy was “creating a formal month-end close process” (Fastly S-1, 2019).

A startup’s version of the control is simpler: whoever books an entry is not its only reviewer, a founder or a fractional finance lead signs the package, and any movement above a set threshold gets a written explanation. Automation belongs in front of the reviewer. In change 5 of Exhibit 3, review by exception halves the review because software clears routine items and passes on the rest; skipping review to hit day 3 leaves the errors for the audit.

Twelve months of delivery days set the date a close can keep

Sign-off sends the package, and its arrival day is the first result a board can check; on US markets, investors price it. Among 2,115 first-time late filers, share prices fell on the late-filing notice, Form 12b-25, even when management promised to meet the grace period (Bartov & Konchitchki, 2017). A startup has no share price, but where the NVCA clause is signed, a missed day breaks a written promise.

Which day a close can keep is a statistical question. Walter Shewhart of Bell Telephone Laboratories called a phenomenon controlled when past experience predicts, within limits, “how the phenomenon may be expected to vary in the future” (Shewhart, 1931). The individuals (XmR) chart built on his work sets those limits from a short run of readings; a close yields 12 a year.

Exhibit 4. The faster close cannot promise the earlier date

	Delivery day, January to December	Mean	Average month-to-month change	Upper limit (mean + 2.66 × change)
Company A	12, 11, 12, 10, 12, 13, 11, 20, 12, 11, 13, 12	12.4	2.7	19.7; August (20) above it
Company B	6, 9, 5, 14, 7, 6, 12, 5, 8, 15, 6, 7	8.3	5.2	22.1; no month above it

Source: hypothetical companies, both promising investors a package within 30 days; calendar days after month end.

The upper limit is what each process will deliver “without fundamental changes” (Wikipedia). Company B closes 4.1 days sooner on average and still cannot promise earlier than day 22. Company A’s August lies above its limit, the mark of an assignable cause such as a stopped bank feed. Fix that cause, and the other 11 months put A’s limit at 15.5, so A can promise day 16 and keep it. B has no single cause to fix: its spread is the process, and only taking waits off its critical path will narrow it.

The record behind the chart is one line a month: the promised day, the actual day and the cause of any gap, the explanation Form 12b-25 demands of a listed company that files late.

Deshmukh and Holzmeier’s managers see no reason to close faster

About 84% of Deshmukh and Holzmeier’s respondents called the length of their close appropriate. The authors concluded: “There’s no real desire among managers to accelerate the closing process further.” They also suggest that speed and the quality of the statements pull against each other.

The SEC reached the same view in 2005, when it kept accelerated filers’ 10-Q deadline at 40 days instead of the scheduled 35. KPMG’s data suggested 35 days could hinder management’s review, and the Commission saw no “net benefit from continuing to accelerate the Form 10-Q” (SEC Release 33-8644, 2005). Where its earlier cuts took effect, earnings quality fell as audits lost time, which Tamara Lambert and co-authors read as lower audit quality (Lambert et al., 2017); that study measures audit time, and a close can be fast without squeezing an audit. At the slow end, 54% of late 10-K filers reported material control weaknesses, against 12% of timely filers (Cao, Chen & Higgs, 2016), and there the weaknesses may cause the delay rather than follow from it.

The German managers have the better of the argument on the target. A seed-stage company with a few hundred transactions a month and a quarterly board gains little from day 3, and forcing three days onto an unchanged process moves error from the calendar into the estimates. Their survey says nothing against the method. Mapping the close, accruing on a schedule and reconciling cash during the month also make a ten-day close hold its date, and they show which numbers are estimates.

Map the close, then shorten the longest chain

1. Draw the close once as a network of tasks, durations and waits, and automate in order of drag; cash matching saves days only when cash is on the critical path.
2. Write an accrual policy before chasing speed: which costs are accrued on day 1, from what (contracts, purchase orders, run rates), and how each true-up is logged.
3. Reconcile cash weekly or daily during the month, and book the reviewer for a fixed day.
4. Record three dates every month: the promised day, the delivery day and the last-entry day, with a cause for every miss. Show them to the board with last quarter’s true-ups. The month-end close checklist and P&L review template below cover the same tasks, grouped by account.
5. Promise a day only once a year of that record sets its limit. To move the limit, look at the critical path first; a part-time controller or a payroll integration may do more than a new tool.

A month-end close is a small project run 12 times a year, and in a small company its critical path runs through waits. Taking the waits off the path turns a ten-day close into a three-day one by estimating sooner, a trade that holds only when the estimates are logged, trued up and reviewed. A company that reports only its delivery day cannot show its board whether they were, or whether that day is one its close can keep.

Sources

• APQC, reported in Wiggins, P. D. (2018). Metric of the month: cycle time for monthly close. CFO.com.
• Bartov, E. and Konchitchki, Y. (2017). SEC filings, regulatory deadlines, and capital market consequences. Accounting Horizons 31(4).
• Bureau of Labor Statistics (2023). Length of pay periods in the Current Employment Statistics survey (February 2023 data).
• Cao, J., Chen, F. and Higgs, J. L. (2016). Late for a very important date: financial reporting and audit implications of late 10-K filings. Review of Accounting Studies 21, 633–671.
• CFO.com (2001). A virtual close: easy as one, two, three? 8 March 2001.
• Cisco Systems, Inc. (2001). Third-quarter earnings release, 8 May 2001 (Exhibit 99.1 to Form 8-K); 2001 Annual Report (Exhibit 13 to Form 10-K).
• Dechow, P. M. and Dichev, I. D. (2002). The quality of accruals and earnings: the role of accrual estimation errors. The Accounting Review 77 (supplement), 35–59.
• Deshmukh, A. V. and Holzmeier, M. (2023). The financial close process: insights from Europe. Strategic Finance, September.
• Fastly, Inc. (2019). Form S-1 registration statement, filed 19 April 2019.
• Ge, W. and McVay, S. E. (2005). The disclosure of material weaknesses in internal control after the Sarbanes-Oxley Act. Accounting Horizons 19(3), 137–158.
• International Auditing and Assurance Standards Board (2019). ISA 315 (Revised 2019), Identifying and Assessing the Risks of Material Misstatement.
• Kelley, J. E. (1961). Critical-path planning and scheduling: mathematical basis. Operations Research 9(3), 296–320.
• Kruze Consulting (2024). Startup Payroll Costs and Tax Budgeting for New Hires.
• Kugel, R. (2019). Finally, a faster close. Ventana Research.
• Lambert, T. A., Jones, K. L., Brazel, J. F. and Showalter, D. S. (2017). Audit time pressure and earnings quality: an examination of accelerated filings. Accounting, Organizations and Society 58, 50–66.
• Ledge (2025). The state of month-end close in 2025: finance team benchmarks and insights.
• National Venture Capital Association (2023). Model Investors’ Rights Agreement, October 2023 revision.
• Sarbanes-Oxley Act of 2002, § 302 (15 U.S.C. § 7241).
• Securities and Exchange Commission. Form 10-Q, general instructions.
• Securities and Exchange Commission (2005). Release 33-8644, Revisions to accelerated filer definition and accelerated deadlines for filing periodic reports.
• Shewhart, W. A. (1931). Economic Control of Quality of Manufactured Product. New York: D. Van Nostrand.
• Ventana Research (2014). Keys to a Fast, Clean Close (white paper sponsored by BlackLine).
• Wikipedia. Critical path method.
• Wikipedia. Shewhart individuals control chart.

The figures come from the surveys, studies, standards and filings listed above, and every close benchmark among them is self-reported. Durations, percentages and control limits without a citation are our calculation from the sources listed; the companies in Exhibits 3 and 4 and the true-up example are hypothetical, and Exhibit 1’s timings come from Exhibit 3. Applying the research on US listed companies’ deadlines, late filings and control weaknesses to private startups is our inference.