Why a 100-Invoice Month Still Takes 8 Days to Close

A 100-invoice month shouldn't take 8 days. It does because the time goes on waiting, review and rework, not volume. Entering 100 bills takes an afternoon. The days go on holding the books open for late bills, chasing other departments, and one reviewer checking everything. Cut the waiting and the rework, not the typing.

How Long Do Other Teams Take to Close?

Longer than they'd like. Ledge, a close-software vendor, surveyed 100 finance professionals at companies from 51 to over 10,000 employees (Ledge). It's a small sample with no split by company size, so treat it as a rough guide.

Half take six days or more. The most common target was 3 to 5 business days. So an 8-day close at a small company isn't unusual. It's just slow.

The same survey names the blockers. The top one, at 56%, is dependency on other departments and regions. Excel came next at 50%, and understaffing at 37%. Ledge's summary: "It's not the final reporting that slows teams down."

Where Do the 8 Days Actually Go?

Figures are illustrative.

Take a US$20M company with a three-person finance team on QuickBooks Online and about 100 vendor bills a month. Here's a typical 8-day close, with the hands-on hours and what each day is really waiting on.

That's 33 hands-on hours, about four working days for one person, stretched across eight. Entering the 100 bills is roughly four of those hours.

Look at the right-hand column. Almost every day ends waiting on someone outside the finance team, or on the one person allowed to review.

Why Doesn't More AP Automation Fix It?

Because it speeds up invoices, and invoices aren't where the days go. Ardent Partners' State of ePayables 2025 puts average invoice processing at 8.2 days. Teams with advanced automation get to 2.9 (Ardent Partners via Medius).

Faster approval helps cutoff, because fewer March bills are still sitting in someone's queue on April 6. But it leaves the review, the department questions and the rework exactly where they were.

The usual fixes miss for the same reason:

• Another hire adds hands to the typing, which wasn't the constraint. They also take months to learn your close, and walk out with that knowledge.
• Close management software shows you, in a nice checklist, that you're waiting. It's built for every company's close, so your exceptions still land on the controller.
• ChatGPT drafts quickly. Then the controller has to review the AI's work as well, with no trail of where its numbers came from.
• An outsourced bookkeeper runs the same process from someone else's desk, on their calendar. That's one more party to wait on.

How to Take Three Days Out

1. Time one close. For each task, log hands-on hours and elapsed days. The gap is your target, not the hours.
2. Fix the cutoff window. Stop entering last month's bills at business day 3. Accrue the rest from evidence, and true it up next month.
3. Ask before month end. Send budget owners "what did you commit to or receive this month?" on the 25th, not the 3rd.
4. Reconcile weekly. Bank and card recs done weekly leave a few days of matching at month end, not a month's worth.
5. Tie out before review. Every schedule should tie to the trial balance before the controller sees it. Review is for judgment, not arithmetic.
6. Review by threshold. The reviewer checks exceptions and lines over a set amount, and samples the rest.
7. Log every correction. Fix the cause before the next close, so the same rework doesn't come back.

Steps 2 and 3 usually matter most, because they attack the waiting directly.

What to Automate vs Keep With a Person

Automate (calculate or draft):

• Bank and card matching, run weekly
• The missing-bill list from vendor history
• Tie-outs of every schedule to the trial balance
• First-draft flux commentary that cites the data, with open questions listed

Keep with a person:

• Whether a cost was incurred and how to estimate it
• Review and sign-off on every schedule and entry
• The answers only department heads have
• What goes in the management pack

How Flux Runs This

We start with your eight days, not a product demo. We map where the close time actually goes, hour by hour, with whoever does the work. That includes the steps nobody wrote down: the clearing account zeroed by hand, the "under $500, code it to general expense" rule. A tool set up from the written procedure breaks on those first.

Then we take on the workstream with the most waiting in it, often accruals or reconciliations. The numbers are calculated in code and tied to your ledger before anything is written. AI reads bills and memos and drafts the reasoning. It never does the math.

What the data can't answer becomes a flagged question with a named owner, sent early. Questions to department heads go out before month end, not on day six.

Your team reviews and signs off. Because every schedule already ties, review is about judgment. We keep each correction, so next month's rework is smaller.

We work from files you choose, with no system access, and a mutual NDA is signed before files are shared. The work gets cheaper for us to run each month. You see that as a shorter close.

FAQ

How many days should the month-end close take? In Ledge's vendor survey of 100 teams, 3 to 5 business days was the most common target. Half of teams take six or more.

Will AP automation shorten our close? Faster bill approval helps cutoff. It won't touch review, rework or waiting on other departments, which is where most of the days go.

What's the quickest fix for a slow close? Set a firm cutoff window and accrue the rest. Then email budget owners before month end instead of after.

Do we need another person to close faster? Usually not first. Time one close and see how much is waiting. More hands don't shorten waiting.