The monthly sequence that turns transactions into reportable numbers — and where the three weeks go.
Period close is the process of finalising a financial period: completing transactions, reconciling sub-ledgers to the general ledger, posting accruals and adjustments, and then locking the period so the numbers cannot change. In an ERP the sequence matters — sub-ledgers close before the general ledger, and inventory before cost of sales can be trusted.
A close that takes three weeks is not a finance problem; it is an operations problem arriving in finance. Goods received but not invoiced, work orders left open, unapplied cash and unreconciled stock all surface at close because that is when someone finally looks. The way to shorten a close is to fix what feeds it, not to work later in the month.
Receiving, despatch and production transactions for the period are completed. Late postings are the main cause of restatement.
Stock movements valued, work orders closed, variances cleared. Cost of sales is not reliable until this is done.
Payables, receivables and cash each reconciled to their control accounts, in sequence.
Goods received not invoiced, prepayments, depreciation and provisions posted.
Reconciliations signed off, the period locked, and reporting produced from a position that can no longer move.
A business closes in eleven working days and treats this as a finance capacity problem. Mapping the close by day shows something different.
Days one to four are spent waiting — for the warehouse to finish posting goods receipts, for the sales team to confirm cut-off, for a bank statement. No accounting work happens; finance is blocked on other departments. Days five to eight are reconciliation, much of it manual spreadsheet comparison that could be automated. Days nine to eleven are review and adjustment, including several journals correcting things found in reconciliation.
The highest-return fix is not in finance at all: a hard operational cut-off at the end of the last working day, enforced by closing the period for operational posting, removes most of days one to four. Automating the three largest reconciliations removes much of five to eight. The recurring adjustments point at specific process faults upstream, each of which is a separate small fix. Four days is a realistic target and almost none of the work to get there is accounting work.
| Term | What it means |
|---|---|
| Cut-off | The line deciding which period a transaction belongs to. Most close errors are cut-off errors. |
| Accrual | Recognising a cost or revenue before its document arrives, so the period is complete. |
| Reconciliation | Proving a ledger balance against an independent source. |
| Period status | Open, closed or permanently closed. Reopening a closed period should be rare and visible. |
| Hard close | A close that cannot be reopened, as distinct from a soft close that can. |
Closing the general ledger before inventory produces a cost of sales figure that then changes.
Differences found at year end cannot be traced. Reconcile monthly while the cause is findable.
They hold value in WIP and distort both inventory and cost of sales until closed.
A period reopened after reporting means the numbers already circulated were wrong. Treat it as an exception requiring approval.
| Metric | What it tells you | What distorts it |
|---|---|---|
| Days to close | The headline. Decompose it by activity before trying to improve it. | |
| Manual journals at close | Each recurring one is an upstream process fault with a name. | |
| Reconciling items ageing | Items carried forward month after month, which are unresolved problems rather than timing. | |
| Post-close adjustments | Corrections after the period was declared closed. A measure of whether the close is real. |
We build and support the systems this page describes. Start with a free working proof of concept — no cost, no commitment.