The Month End Close Process: A Calendar, an Owner for Every Step, and the Places It Jams

Month-end close is late again. It's day six, the controller has a spreadsheet called close_tracker_v4_FINAL open on one screen and Slack on the other — every message says some version of "waiting on someone else."

Payroll is waiting on headcount confirmation. Revenue is waiting on legal to clear a credit note. Accounts payable is waiting on goods receipts nobody entered while the usual person was on leave. And the bank reconciliation is off by $2,847 — a processor fee that changed rate on the 14th, though nobody will find that out until 4 p.m.

None of this is a competence problem. It's a sequencing problem: too many steps blocked by other steps, no visibility into what's actually holding things up, and a tracker that shows status but not dependency. Six people are busy, and the close still isn't moving.

What the month end close process has to produce

A month is closed when three statements are all true. The subledgers agree with the general ledger. Every balance sheet account has a reconciliation supporting it. And the period is locked, so no further entries can be posted to it without a documented exception.

The third one is the one teams skip.

A close that isn't locked can be silently amended, which means last month's reported numbers and last month's current numbers can differ without anyone deciding they should. Once that's possible, the close has produced a draft rather than a result, and every downstream conversation about the figures carries an invisible asterisk.

Locking is also what makes the timetable meaningful.

If entries can still land on day nine, day five was never a deadline. It was a suggestion with a date on it, and everyone downstream learns that within a quarter.

The month end close process, day by day

Days here are working days after the period end, written as D+1, D+2 and so on. Groups with subsidiaries add consolidation days at the end, and a company holding inventory adds a day around counts.

Two things about that table are worth arguing over inside your own company.

The first is that every row has exactly one owner. Not a team.

A person. A step owned by "finance" is a step that gets started when someone happens to notice it hasn't been, which is usually on the day it becomes urgent.

The second is that D+3 is a single day for all reconciliations. That only works if they're prepared during the month rather than begun on D+3, with the bank and processor accounts reconciled weekly and the control accounts checked as the subledgers move.

That single choice is what separates a close that finishes from one that drags. It isn't a tooling difference, and no software fixes it.

Cutoffs decide which month a cost lands in

The recurring argument in every close is an invoice dated the 29th, received on the 3rd, for something delivered on the 27th.

It belongs to the earlier month.

The event that creates the cost is delivery. Not invoicing, and not receipt of the paperwork. This is the entire reason accruals exist: so that a supplier's administrative speed can't move your reported cost base, and so that two months of costs don't land in one because a vendor changed billing systems.

Three cutoffs need a stated hour rather than a stated day:

Receiving. Goods and services accepted after the cutoff belong to next month, and the receipt record has to carry the date of actual delivery, not the date somebody got round to entering it.

Expenses. Claims submitted after the cutoff go in next month. Making an exception for one senior person's late claim is how the rule dies.

Billing. Revenue events after the cutoff belong to next month, and the exception process for a large late deal needs to exist in writing before the month you need it.

The cutoff failure that causes the most trouble isn't lateness.

It's a receipt entered on the 3rd carrying the 3rd as its date. That quietly moves a cost into the wrong period and produces a reconciliation difference nobody can explain, because the underlying record now looks perfectly correct.

Four accruals you will need every month

An accrual is a cost you've incurred and not yet been billed for. Four of them recur so reliably that they should be standing entries with a documented calculation, not a fresh judgement each month.

Goods received not invoiced. Anything with a goods receipt and no matching invoice. This one is mechanical if the receiving discipline in your accounts payable process holds, and guesswork if it doesn't.

Unbilled revenue. Work delivered and not yet invoiced. The mirror image of the above, and the one auditors examine hardest because it's where optimism lives.

Payroll and related costs. Where the pay period doesn't align with the calendar month, the days between the last pay date and month end are a real cost with no transaction behind them.

Usage-based services. Cloud, payment processing, telephony. Billed in arrears, often on a cycle that doesn't match the month, and frequently large enough to matter.

Each one needs a written basis: what data feeds it, who calculates it, and whether it reverses automatically next month.

An accrual with no documented basis becomes untouchable. It sits on the balance sheet for two years because nobody remembers what it was for, and nobody wants to be the person who wrote off something that turned out to be real.

Subledgers close before the general ledger

AP, AR, payroll, fixed assets, and inventory each maintain their own detail, and each must agree to a control account in the general ledger before the general ledger can be trusted.

The order matters. If AP is still open while someone is reconciling the AP control account, the reconciliation is being performed against a moving target, and it will need to be done again. Freezing the subledger first isn't bureaucracy, it's what makes the reconciliation valid.

A useful test: pick the AP control account and the AP aged listing and compare the totals.

If they don't match and nobody knows why, the close hasn't started yet, whatever the tracker says.

Reversing journals, and what they hide

Accruals reverse in the following period, which is correct and also the source of a specific recurring problem.

When an accrual reverses and the actual invoice arrives, the two should roughly cancel. When they don't, the difference tells you something: your accrual basis is wrong, or the invoice is wrong, or the cost belonged somewhere else. Most teams post the reversal automatically and never look at the difference, which throws away the one feedback signal the process generates.

Tracking accrual accuracy over three or four months is a small piece of work that pays back quickly, because the pattern is usually consistent rather than random: the same categories run light, for the same structural reason, every month.

A category accrued 30% light four months running isn't rounding. It's a basis that needs rewriting.

The reconciliations that gate the close

Not every reconciliation blocks the close, and treating them all as equally urgent is why day three overruns. Three categories genuinely gate it.

Cash. Every bank account, plus every payment processor balance. A processor is not a bank: the balance moves on a settlement cycle, arrives net of fees, and includes reversals that relate to earlier periods. That mismatch has its own mechanics, set out in PSP settlement reconciliation.

Control accounts. AP, AR, payroll, and tax. These are the accounts where a difference means the detail and the summary disagree, which is a data integrity problem rather than a timing one.

Intercompany. Balances between entities must eliminate exactly. A difference here survives consolidation and shows up in the group accounts, which is a much more expensive place to find it.

Where part of the balance sheet sits on a blockchain, the same gating logic applies and the matching identifiers are different: there's no payment reference, and the transaction hash does the work a statement line ID would do. The mechanics of that match are in crypto reconciliation, and the record-keeping frame underneath it is in crypto bookkeeping.

Flux analysis answers one question

Compare each line to the prior month and to budget, and explain anything that moved more than a stated threshold. That's flux analysis, and its value isn't the explanation, it's the question it forces: does anybody understand this number?

A cost centre that moved 40% with a confident explanation is fine.

A cost centre that moved 4% with nobody able to say why is the one to look at. A small net movement can hide two large offsetting errors, and the netting is exactly what makes it invisible on the report.

Set the threshold as both a percentage and an absolute amount. A 60% movement on a $900 line is noise. A 3% movement on a $2.4 million line is $72,000 and deserves a sentence.

Hard close, soft close, and the month end close checklist that fits each

A hard close does everything: full accruals, full reconciliations, full review, locked period. Required at year end, and for most companies at each quarter. The financial close checklist is the same document in each case, with different rows marked optional.

A soft close does the material items and accepts known imprecision in the rest. Appropriate for months where the numbers feed internal management reporting only, provided the shortcuts are documented rather than improvised.

A continuous close moves work into the month: reconciliations performed weekly, accruals maintained as standing entries, cutoffs enforced by system controls rather than by reminder emails. It isn't a different close, it's the same close with less of it happening in the first week.

The choice is a real one. The mistake is drifting between them without deciding.

A team that runs soft closes for eleven months and then attempts a hard year end discovers in January exactly how much imprecision accumulated, usually in the form of balance sheet accounts that have not been looked at since the previous audit.

Where the month end close process actually jams

Four bottlenecks account for most overruns, and each has a specific fix.

The fourth is the one finance teams resist.

A stated materiality threshold, applied consistently and written down, is defensible to an auditor and to a board. Chasing every difference to zero is not more rigorous. It moves rigour to where it produces nothing, and takes it from where it would have.

When part of the ledger settles on-chain

Companies paying suppliers or contractors in stablecoins have a close that looks the same in structure and differs in three specific places.

Cash reconciliation covers wallets as well as bank accounts, and wallet balances move on network confirmation rather than on a banking day. Network fees are an operating cost that has to be split from the payment amount rather than netted into it. And the rate applied to convert a token amount into the reporting currency needs a written policy, because the alternative is a different defensible answer every month.

The expense side of that has its own month-end sequence, covered in crypto expense management. Platforms built for this, VaultNow among them, keep custodial and external wallets on one dashboard alongside payouts and invoicing, which is what turns a wallet reconciliation into a normal day-three task rather than a research project.

Frequently asked questions

How long should the month end close process take?

Long enough to lock the period with every balance sheet account reconciled, and the binding constraint is almost never company size. It is how much of the reconciliation work was left until after the period end. Day counts circulate widely as benchmarks, but the credible research sits behind subscription paywalls and the figures repeated in vendor articles are second-hand, so treat any specific number quoted without a source as decoration rather than a target.

What is the difference between a hard close and a soft close?

A hard close performs full accruals, reconciliations, and review, and ends with the period locked. A soft close addresses material items and accepts documented imprecision elsewhere, which suits months where the output feeds internal reporting only. Year end and most quarter ends need a hard close.

Which period does an invoice belong to?

The period in which the goods or services were received, regardless of the invoice date or the date the document reached you. The accrual for goods received not invoiced exists to hold that cost in the correct period until the invoice arrives, and it also stops days payable outstanding swinging on a supplier's admin speed.

What is flux analysis in the month end close process?

Comparing each line against the prior period and against budget, and explaining any movement above a stated threshold. Set the threshold as both a percentage and an absolute amount, so that small percentages on large balances still get examined.

Why do subledgers need to close before the general ledger?

Because reconciling a control account while its subledger is still accepting entries means reconciling against a moving total, and the work has to be repeated. Freezing AP, AR, and payroll first is what makes the general ledger reconciliation valid.

What should a month end close checklist contain?

Every step, its owner by name, the day it is due, and what it depends on. The dependency column is the one most versions of a financial close checklist omit, and it is the one that tells you which delay is actually holding up the close.

How do you speed up the close without cutting corners?

Move work into the month. Reconcile weekly rather than at period end, keep recurring accruals as standing entries with documented bases, enforce cutoffs through system controls, and set a materiality threshold so that immaterial differences stop consuming days.

The month end close process runs late for structural reasons, not because people are slow: work that could have happened during the month is queued into the first week, one person's step blocks four others, and there's no agreed point at which a small difference stops being worth chasing.

If some of your payments settle on-chain, the wallet side of the reconciliation is the part that most often has no owner. VaultNow puts custodial and external wallets, payouts, and invoicing on a single dashboard, so the wallet leg closes on the same day as the bank leg.

General information, current as at 31 August 2026. Not accounting, tax, or legal advice.