What the month-end close really is (an owner's guide)

What is the month-end close process, and why does it exist?

The month-end close is the sequence of steps that turns a running transaction log into a fixed, defensible view of one period.

That is the whole idea. During the month your books are a live feed: bank transactions arriving, invoices going out, bills being entered, a card charge landing three days after the meal. None of that is wrong, and none of it is final either. The close is where you stop the feed, check it against reality, add the things reality knows about that the ledger does not, and then say: this is what happened in June.

Owners often assume the close is a formality the bookkeeper performs at some point. It is not a formality. It is the only moment in the month when somebody deliberately asks whether the numbers are true. An unclosed ledger is a ledger nobody can act on, because if the period is still open, every number in it is provisional. Revenue can move because a late invoice gets backdated. Margin can move because a supplier bill arrives on the 12th and gets posted to the prior month. You can look at the same June profit figure twice in three weeks and get two different answers, which is worse than having no answer, because the first one already went into a decision.

Closing puts a fence around the period. After the fence goes up, the number you looked at last week is the number you will see next year.

What actually happens between the last day of the month and a finished close?

Four kinds of work, roughly in order.

Cut-off and collection. Every source of transactions has to be complete for the period. Bank and card statements, payroll runs, sales system exports, supplier invoices that arrived by email and are still sitting in an inbox. You cannot reconcile against a source you have not received.

Reconciliation. Each balance sheet account is agreed to something external or independently verifiable. Bank accounts to statements. Credit cards to statements. Accounts receivable to the aging report. Loans to the lender's amortisation schedule. Reconciliation is not "the balance looks about right"; it is "here is the difference, and here is what each item in the difference is."

Accruals and adjusting entries. Reality does not arrive on a monthly schedule, so you record the parts of it the ledger has missed. Depreciation. Prepaid insurance spread across the year it covers. The interco management fee. Revenue earned in June but invoiced in July. Payroll days that straddle the month end.

Review and lock. Somebody who did not prepare the work looks at it, asks about anything that moved oddly, and then the period is closed in the accounting system so nothing new can post to it without a deliberate decision.

What does "closed" actually mean?

A period is closed when no further entries can be posted to it without an explicit, recorded decision to reopen it.

That word "recorded" matters. Plenty of small groups think they close, but in practice the prior month stays open all year and entries drift backwards into it. If you cannot tell me who reopened April and why, April was never closed. It was just old.

How long should a close take?

APQC's Open Standards Benchmarking work, covering roughly 2,300 organizations, puts the median monthly close at 6.4 calendar days, the top quartile at 4.8 days, and the bottom quartile at 10 days or more.

Those numbers are a useful reality check in both directions. If your books are final on business day five, you are doing well and you should stop apologising for it. If you are getting a PDF on day 15, you are not unusual, but you are in the bottom quartile of a benchmark that includes companies vastly more complicated than yours.

The honest caveat: those figures cover organizations of every size, most of them running one set of books. A group of eight entities has more work to do than the benchmark's median respondent. It also, usually, has far fewer people doing it. That is the whole tension, and it is worth understanding why a close stretches to 15 days before you try to compress it.

Where the close breaks when you own several businesses

The single-entity close is a checklist. The multi-entity close is a chain, and chains fail at the joints.

Intuit's Enterprise Technology Benchmark found that 76% of multi-entity firms say their technology struggled when they added entities, and 64% say the close takes too long. That is not a story about bad bookkeepers. It is a story about a process designed for one company being run eight times in parallel by tools that only ever expected one.

The joints are the interesting part: intercompany balances that have to agree between two sets of books, shared costs allocated across entities on some basis somebody chose two years ago, a management company that pays for things on behalf of the operating businesses, and a consolidated view that cannot be produced until the last entity is done.

A worked example: a group of nine entities

Say you own nine entities. One holdco, one management company, five operating businesses, two property companies.

The management company pays a single insurance premium covering all seven trading entities and recharges it monthly. Entity A bills Entity B rent. The holdco charges a management fee to each operating business. Payroll runs centrally out of the management company and is allocated by headcount.

Now count the dependencies. The insurance recharge cannot be posted until the premium is entered. Entity B cannot close until the rent invoice from Entity A exists. The consolidated numbers cannot exist until all nine are reconciled and every intercompany pair agrees. And the allocation basis for payroll depends on a headcount file that lives with whoever runs HR.

Nothing in that list is hard. There are just twenty or thirty small waits, and each one begins only when somebody else has finished. That is how a close that contains four days of actual work takes three weeks of calendar.

What does a rushed close cost?

Speed without checking produces errors, and errors in a closed period are expensive to unwind.

Gartner's 2024 survey of 497 controllers and chief accounting officers found that 18% of accountants make financial errors at least daily and 59% make several errors a month. Most are caught. The ones that are not tend to surface later, in a lender conversation or a tax filing, at exactly the moment you would prefer them not to.

The point is not that errors are shameful. The point is that a close is where they are supposed to be caught, and a close that is mostly a rush to publish something is not doing that job.

Can QuickBooks or Xero do this?

For one entity, yes, genuinely well. Both have solid bank feeds, real reconciliation screens, recurring journal templates, and a period-lock function that works if you turn it on. A single company with a competent bookkeeper can run a clean five-day close in either one, and should.

The gap is the group. Neither was built to reconcile Entity A's intercompany receivable against Entity B's payable, to eliminate those balances on consolidation, or to give you one status view across nine files that tells you which entity is holding up the group. What owners actually do is export each file to a spreadsheet and rebuild the group by hand every month. That works until it does not, and it fails silently, which is the part that should worry you. It is worth reading about where multi-entity consolidation actually breaks before you accept the spreadsheet as permanent.

What should a finished close package contain?

At minimum: a balance sheet and profit and loss per entity, a consolidated view, a cash summary, the reconciliations supporting each material balance sheet account, a list of the journal entries booked in the period with their support, and a short note on anything unusual.

That last item is the one most packages skip and the one owners actually read. Three sentences explaining why gross margin moved four points is worth more than another tab of numbers.

How do you tell whether your close is working?

Four questions, answered honestly.

How many calendar days after month end do you have final numbers? Do the numbers change after you receive them? Can you see, in writing, what supports each balance sheet account? And if your bookkeeper disappeared next Tuesday, could someone else run next month's close from what is documented?

If the fourth question makes you uncomfortable, that is the one to work on first. It is also closely related to your control environment, because a close that lives in one person's head is a close nobody can review.

Where to start

If your close has never really existed as a defined process, do not start by trying to compress it. Start by writing down what actually happens now: every entity, every source of data, every person who has to do something, and the order it happens in. Most owners find four or five waits that could run in parallel, and one or two that could disappear entirely.

Then pick a target date and hold it. A close that lands on business day 3 every month is worth more than one that occasionally lands on day 4.

That fixed cadence is the idea behind how cruisr runs a close: it sits on the QuickBooks Online or Xero files you already have, works on an "AI prepares, humans approve" basis so the books are current nightly rather than rebuilt monthly, routes exceptions to a human team, and delivers a finished close package by business day 3. Nothing migrates, the files stay in your name, and cruisr never moves or holds money.

If you want a second pair of eyes on where your own close is actually losing days, get in touch. cruisr runs a free close diagnostic on your existing QuickBooks or Xero files and comes back within 48 hours with a 30-minute readout of what it found.