The close itself — reconciliations, accruals, reports — is a day's work in most small businesses. Ask anyone who does it and they'll tell you it takes three or four.
The difference is almost never the close. It's the backlog of purchase invoices that arrived during the month and are being dealt with now, at the worst possible time.
The shape of the problem
Invoices arrive continuously. They get entered in a batch. The gap between those two things is where the trouble accumulates.
If you enter as you go, the close starts with a purchase ledger that's already current. If you enter monthly — which most small teams do, because there's always something more urgent — the close starts with a stack, and everything else waits behind it.
The stack is also worse than the same documents would have been earlier:
Queries have gone cold. An invoice you can't code because you don't recognise the purchase is a two-minute question in week one. In week five, the person who ordered it has forgotten, and it's a conversation.
Missing documents surface late. The supplier who never sent March's invoice is invisible until you reconcile their statement — and by then you're inside the close.
Cut-off gets ambiguous. Goods received in March, invoiced in April, entered in May. Deciding what belongs where is fiddly, and it's being decided under time pressure by someone who wants to go home.
Where the days actually go
Worth separating the parts, because they have different fixes.
Entering the backlog. Pure volume. Fifty invoices at four minutes is over three hours before anything else starts.
Chasing approvals. Invoices needing sign-off go out during the close, and approvers are busy people who aren't thinking about your deadline. This is often the single largest source of elapsed time — hours of waiting rather than hours of work.
Statement reconciliation. Comparing supplier statements to your ledger to find what's missing. Slow, and it's where the late surprises come from.
Accruals for what hasn't arrived. The invoice that hasn't come yet still needs to be in the right period. Every one is a judgement call.
Fixing coding errors. Entry done at speed produces mistakes, and the review pass finds them at exactly the moment there's no time to fix them.
What actually shortens it
Move entry out of the close. The largest single win, and the hardest to sustain by willpower alone. Weekly entry sessions turn a three-hour block into four forty-minute ones that happen when there's no deadline attached.
Get approvals moving before month-end. Approval is elapsed time, not work time — the constraint is other people's attention. Sending approval requests on the 25th rather than the 2nd costs you nothing and removes days of waiting.
Reconcile statements weekly. Missing invoices found on the 10th are a routine email. Found on the 3rd of the following month, they're a problem.
Keep a standing accruals list. The same recurring costs need accruing every month. Maintaining the list continuously means the close is a review rather than a reconstruction.
Set supplier defaults. Coding decisions made once per supplier rather than once per invoice. Half an hour of setup, permanent return.
The uncomfortable bit
Most of these amount to the same instruction: do the work earlier.
Which is true, and it's also why they don't stick. Nobody enters invoices late because they think it's a good idea. They do it because on any given Tuesday there's something more pressing, and the invoices aren't due yet.
The advice "be more disciplined" has been available to everyone the whole time.
That's the honest case for automating entry — not that it's faster in the moment, but that it removes the thing you can defer. A document that's read when it arrives can't be deferred to month-end, because there's nothing left to defer.
The rest of the close is still yours. Judgement, accruals, review — none of that disappears. But it starts from a current ledger, which is a different job from starting with a stack.
A reasonable target
If you're closing in four days, ask which of those days is genuinely the close and which is catch-up. Most people find it's one or two of each.
Getting the catch-up out of the close is worth more than optimising the close itself. It's also the part that doesn't require anyone to work faster.
Cribble reads each document as it arrives, so the entry work doesn't accumulate for month-end. The close still needs your judgement — it just doesn't need to start with a stack.