How a Continuous Close Works in Bookkeeping, and What It Demands of You

The month-end close is a batch job. Work piles up for four weeks, then a small window of days absorbs all of it. A continuous close spreads that same work across the month, so that at any point the ledger is close to final. This article covers the mechanics, because the idea only works if the daily habits are real.

What does "close" actually mean here?

A close is not a single event. It is a set of assertions: every transaction is recorded, every account is supported, cut-off is correct, and the trial balance can be trusted enough to report from. Month-end compresses those assertions into a deadline. A continuous close makes each assertion true as you go.

The end state is the same. You still need a reconciled balance sheet, a reviewed profit and loss, and working papers someone could pick up cold. The difference is sequence, not standard.

Which tasks move from month-end to daily?

The core shift is that anything driven by a feed becomes a same-day task rather than a catch-up task. In practice that means:

• Reviewing imported bank and card transactions daily, and coding them as they land.
• Matching payments to open invoices and bills as cash moves, not at month-end.
• Reviewing newly created bills and invoices for correct account, tax code, and period.
• Posting recurring accruals and prepaid amortisation on a schedule, so nothing waits for a human to remember it.

Notice what is absent: judgement-heavy work. Estimates, provisions, and unusual items still need a person, and they still cluster at period end.

What stays a period-end task?

Some work cannot be continuous by nature. Depreciation runs on a period convention. Accruals for unbilled costs need a cut-off date to be meaningful. Inventory counts, payroll true-ups, and the accountant's adjusting entries all belong to the period boundary.

A sensible split is to automate the mechanical and defer the judgemental. If a task is rule-driven and evidence-backed, run it daily. If it depends on an estimate or a boundary, keep it at close.

Where does the evidence trail change?

In a batch close, the working paper is assembled after the fact. In a continuous close, evidence is created at the moment of coding. Each transaction should carry its own support: the attachment, the matching payment, the approval.

That changes what review looks like. Instead of re-performing a month of work in a weekend, you sample. You review exceptions, new vendors, unusual amounts, and anything the rules did not catch. The working paper becomes a log of what was reviewed and cleared, not a reconstruction.

What are the real risks?

The main risk is false confidence. A ledger that looks current can still be wrong: a payment matched to the wrong invoice, revenue recognised in the wrong period, a duplicate supplier quietly doubling up. Speed of recording says nothing about correctness of classification.

The second risk is drift in cut-off. Daily coding makes it easy to post things in the period they arrive rather than the period they belong to. Accrual accounting still governs. Goods received before period end belong in that period, invoice or no invoice.

How do you run a continuous close well?

We would suggest four disciplines. Reconcile the high-volume accounts weekly at most; daily if the flow justifies it. Keep a standing close checklist with owners and dates, so nothing depends on memory. Review exceptions on a fixed rhythm, and record what you cleared. And hold the judgement calls to a short, scheduled period-end session, so estimation does not get scattered and rushed.

Done well, the month-end meeting becomes a confirmation rather than a scramble. That is the whole point: the close stops being an event and becomes a state of the books.