Month-end close checklist

Why closes run late (it is rarely the accounting)

Ask a finance team why last month took nine days and the answer is almost never "the journals were difficult". It is that three supplier invoices had not arrived, the credit card statement was not downloaded, nobody could explain a difference in the clearing account, and one person held four of the tasks.

In other words, the delays are logistical and structural. Waiting on documents. Waiting on people. Sequential work that could have run in parallel. Investigation of a difference that would have taken ten minutes if cash had been reconciled weekly instead of monthly.

That is good news, because logistics can be fixed cheaply. Most of the improvement in a close comes from moving work earlier and removing the wait states — not from working faster during the close itself.

The close timeline at a glance

Adjust the day numbers to your size — a sole trader does this in an afternoon, a mid-sized company across a week — but keep the order. Doing accruals before cash is reconciled means redoing them, and reviewing before the balance sheet is reconciled means reviewing numbers that are still going to move.

Before the month ends — the week that saves the close

The most valuable close work happens before the month is over, and it is almost entirely about removing wait states.

• Chase missing supplier invoices in the final week, not after the month closes.
• Confirm which accounts have no live feed and schedule their statement downloads for the first working day.
• Draft recurring journals — depreciation, rent, insurance, subscriptions — so they only need reviewing.
• Check that the previous month's reconciling items have actually cleared rather than rolling forward again.
• Warn whoever owns a step that their input is due, especially outside finance.

One habit dominates all of these: reconcile cash weekly during the month. A month reconciled in four weekly passes leaves a close with almost nothing to reconcile, and any difference is found while everybody still remembers the transaction that caused it.

Days 1–2 — Cash first, always

Reconcile every bank account, every credit card, every payment processor and every petty-cash float to its statement. Not approximately — transaction by transaction, with the closing balance agreeing exactly.

Cash is first because everything downstream depends on it. Payables cut-off is meaningless if payments are missing; accruals are guesswork if you do not know what was actually paid; profit is unverifiable. If cash does not reconcile, stop and find out why rather than continuing and hoping the difference explains itself.

Treat payment processors as clearing accounts rather than as income. Gross sales, platform fees and refunds each need separating, and the payout that lands in the bank is the net residue — see processor statement extractionfor the mechanics. Loans need reconciling against the lender's own statement too, with interest and principal split correctly rather than lumped together.

flowparse.io

The statement bottleneck nobody plans for

Here is the step that quietly sets the length of a great many closes: getting transactions out of the accounts that have no feed.

Feeds cover the main accounts and rarely all of them. The foreign account, the older company card, the account at the small institution, the loan, the merchant processor, an account closed mid-period — those arrive as PDFs. Until they are transactions, the cash step cannot finish, and because cash is first, nothing else can finish either.

Two habits fix it. First, download those statements the day they are issued and keep them in one place named by account and period — that removes the scavenger hunt. Second, convert them in bulk rather than retyping: statement conversion turns each PDF into typed, signed transactions and a QBO, Xero CSV or Excel file your ledger imports.

Insist on one thing from whatever tool you use: proof that nothing was dropped. A statement extracted with three rows missing looks perfect and reconciles to nothing — the balance checkcompares the extracted transactions against the statement's own opening and closing balance, per account, and fails loudly when they disagree. That check is what keeps the close from starting on quietly incomplete data.

flowparse.io

Days 2–3 — Payables and cut-off

The payables step is mostly about cut-off: making sure every cost belonging to the month is in the month, and nothing else is.

Post every supplier invoice relating to the period, including ones that arrived after month end for work done before it. Reconcile supplier statements against your ledger to find invoices you never received — that is the single most effective way to catch a missing cost, and it is why capture tools price supplier-statement reconciliation as a distinct feature. Check for duplicates deliberately: the same invoice chased twice by email is paid twice more often than anyone likes to admit.

Then handle the ones with no invoice at all. Regular costs where the bill has not arrived get accrued in the accruals step rather than ignored — and if you are consolidating a pile of purchase invoices into a register to work through them, merging them into one Excel register is the fastest route, with invoice extraction doing the reading.

flowparse.io

Days 3–4 — Receivables

Mirror image of payables. Every sale delivered in the month should be invoiced in the month, receipts should be applied to the right invoices, and the aged debtors list should be reviewed rather than filed.

Two things reliably hide here. Unapplied receipts — cash received but not matched to an invoice — overstate debtors and understate collection, and they accumulate quietly. And revenue cut-off: work delivered on the last day of the month but invoiced on the second of the next belongs to the month it was delivered, not the month it was billed.

Matching is the part that automates well: fuzzy payee names, part payments, one payment covering several invoices, small FX differences. See payment matching and invoice reconciliation for how that is handled at volume, and unpaid invoice detection for finding what has not been collected.

Days 4–5 — Accruals, prepayments and payroll

This is where the month stops being a cash record and becomes a picture of performance.

Keep the calculation with the journal rather than in someone's head. An accrual whose basis nobody can reconstruct is the reason balances survive for years after the underlying cost disappeared — and it is the first thing an auditor asks about.

Days 5–6 — Reconcile the whole balance sheet

Profit gets the attention; the balance sheet is where errors hide. Every balance-sheet account should be reconciled to something external or independently calculable, not simply accepted because it looks familiar.

• Bank and card accounts — to statements, already done in the cash step.
• Payment processor and merchant clearing accounts — to the processor's own statement.
• Loans and finance leases — to the lender's statement, split between interest and principal.
• VAT or sales-tax control accounts — to the return and the underlying transactions.
• Payroll control accounts — to the payroll report and what was actually paid.
• Debtors and creditors — to the aged sub-ledgers, agreeing to the nominal balance.
• Suspense and uncategorised — to zero. Not small. Zero.

The suspense rule is worth being rigid about. A balance parked there is an unanswered question, and unanswered questions compound: by the third month nobody remembers the transaction well enough to answer it, and it becomes a year-end problem instead of a ten-minute one.

flowparse.io

Days 6–7 — Review for plausibility, not arithmetic

By this point the numbers add up. The review is about whether they are believable — which is a different skill and the one that catches systematic errors.

Then somebody other than the preparer signs it off, with the reconciliations attached. That separation is the control that catches both honest mistakes and the rarer deliberate ones, and it costs nothing to implement.

Reporting and sign-off

A close that produces a correct ledger and no explanation has done half the job. The management pack should carry the numbers, the comparison, and a short commentary on what changed and why — three paragraphs from someone who understands the business is worth more than twenty pages of tables.

Issue it on a fixed day. A pack that arrives on the eighth working day every month gets used in decisions; one that arrives somewhere between the fifth and the fifteenth gets ignored, because nobody can build a meeting around it.

Then log what went wrong this month while it is fresh — the invoice that arrived late, the account nobody could reconcile, the person who was on holiday. That log is what turns next month's close from a repeat performance into a shorter one.

The full checklist

What to automate first

Automate in the order of the critical path, not in the order of what is easiest to demo.

First: getting data in.Statement conversion for the non-feed accounts and document capture for purchase invoices. This is almost always the longest wait in the close and the least valuable use of anyone's time.

Second: matching. Payments to invoices, statement lines to ledger entries. Machines are good at fuzzy matching and humans are slow at it — see the reconciliation engine.

Third: recurring journals. Anything that repeats with a predictable calculation should be a template, reviewed rather than re-derived. What should not be automated is the review itself — that is the part the automation exists to make time for. A fuller treatment of the automate/keep-human split is in bookkeeping with AI.

Evidence: what the close should leave behind

A close is not only a set of numbers; it is the reason to believe them. Three months later, someone will ask how a balance was arrived at, and the answer should take a minute rather than a morning.

Keep, per account: the statement or third-party document, the reconciliation showing agreement, and a note on any reconciling item. Keep, per material journal: the calculation and its basis. And keep traceability from the ledger back to the document — if you consolidated several statements into one dataset, a source-file reference on every row is what makes a single figure explicable.

Worth stating plainly: an extraction tool is not an archive. FlowParse deletes the original PDF immediately after extraction, so retention of the source documents for whatever period your jurisdiction requires remains your responsibility — a point covered further in the audit preparation guide.

How to shorten the close

Measure before you optimise. Time each step for two months and the bottleneck will be obvious — it is nearly always waiting for documents or reconciling cash, and it is rarely the step people assume.

Then apply the four levers in order: move work earlier (weekly cash reconciliation, pre-close document chasing), remove wait states (statements downloaded on issue, one intake route per document type), parallelise what does not depend on cash (receivables review, payroll preparation), and automate transcription. Those four routinely take a nine-day close to five without touching the quality of the work.

What does not shorten a close is cutting review. That defers the work to whoever finds the error later, usually at year-end, usually at several times the cost — and it removes exactly the step that makes the numbers worth producing.

flowparse.io

The small-business version

Not every business needs a seven-day process with named owners. For a sole trader or a small company the same sequence compresses into a single session, and the value is in doing it monthly rather than in doing it elaborately.

Reconcile the bank and card. Convert any statement that is not in a feed and check it against its closing balance. Post the purchase documents you have and note the ones you do not. Accrue anything material and recurring. Look at the profit figure and ask whether it matches your sense of the month. Then stop.

An hour a month, done consistently, is what prevents the catch-up job that later takes three days — and if you are already behind, the catch-up guide is the way back.

Closing for many clients at once

Practices have a different problem: the same close, eighty times, with different levels of client cooperation. Three things make it manageable.

Standardise the checklist so no client gets a bespoke process. Stagger deadlines across the month rather than having every client due on the fifth, which just guarantees a bottleneck. And batch the mechanical work — convert every client's statements in one sitting rather than switching context per client, which is where batch processing and the accountants' workflow earn their place.

Then track, per client, which step blocks the close most often. It is usually the same one for the same client every month, and naming it is the first step to fixing it — often by changing what the client has to send rather than what your team has to do.

A worked example: a five-day close

Abstract sequences are easy to agree with, so here is a real shape. A services company with two bank accounts, one company card, a payment processor and about 300 transactions a month, closing on the fifth working day.

Final week of the previous month. The bookkeeper emails three suppliers whose invoices are always late, downloads the card statement the day it is issued, and drafts the depreciation, rent and insurance journals. Cash has been reconciled weekly, so only the last few days remain.

Day 1. Both bank accounts reconcile within an hour because most of the month was already done. The card statement — a PDF, no feed — is converted, checked against its printed closing balance and imported; the processor statement is reconciled to the clearing account, separating gross sales, fees and refunds. Cash is finished by lunchtime.

Day 2. Supplier invoices are posted, two supplier statements reconciled, one duplicate caught before payment. Two costs have no invoice yet and are noted for accrual. Day 3. Sales invoiced, receipts applied, one unapplied receipt from last month resolved, aged debt reviewed and two chasers sent.

Day 4. Accruals, prepayments, depreciation and payroll posted from the drafted templates, each with its calculation attached. The balance sheet is reconciled account by account; suspense is already zero because nothing was parked there during the month.

Day 5. Variance review against last month and budget throws up one question — a cost ratio moved three points — which turns out to be a supplier price rise rather than an error. The owner reviews and signs off, the pack goes out with three paragraphs of commentary, and the close log records that the card statement was again the only thing holding up day one.

Nothing in that week is clever. It is early preparation, weekly reconciliation and one converted statement — and it is the difference between five days and nine.

Common mistakes

Frequently asked questions

Keep reading