Guide 02 · Process
Closing the books in five business days
A late close is almost never a capacity problem. It is a sequencing problem: the team is doing step six while step two is still open, so step six gets done twice. Put the chain in the only order that works and the calendar stops being the constraint.
The close, day by day
A dependency chain, not a checklist — each day's work depends on the one before it.
The close is a chain, not a list
Every close task has a prerequisite, and the expensive mistake is starting a task whose prerequisite is still moving. If cash is not reconciled, your revenue number is provisional. If revenue is provisional, the accrual is provisional. If the accrual is provisional, the P&L you just reviewed with an owner is provisional — and you will review it again.
Work done on a moving input is work you will redo. That, not headcount, is where the second and third week of a slow close actually goes.
Revenue real → AR and adjustments are real
AR real → accruals can be set
Accruals set → intercompany can be eliminated
Intercompany eliminated → consolidation ties
Consolidation ties → variance analysis means something
What must be true before day one
Most of a five-day close is won in the three weeks before it starts. These are continuous, not close-week, activities:
- Bank and card feeds are current to yesterday, every day — not pulled in a batch on day two
- AP is entered on receipt, with a cut-off everyone knows and a hard stop date
- Payroll is reconciled each run, not reconstructed at month end from three registers
- Recurring journal entries (depreciation, amortisation, rent straight-lining, allocations) are templated and scheduled, not retyped
- The chart of accounts has not changed mid-month without the close checklist changing with it
- Every balance sheet account has a named owner (see guide 07)
If your close takes three weeks, ask how much of week one is spent gathering things that could have been gathered in real time. In most groups it is the majority of week one.
Day by day
Day one — cut off and tie cash
Close the subledgers. AP cut-off enforced, AR batches through the last business day posted, payroll accrued through month end. Then reconcile every cash and card account to the three-way proof and post the entries the reconciliation generates. Nothing else starts until cash is signed.
Day two — revenue, AR and the clearing accounts
Bridge the practice management or billing system to the ledger: gross production, adjustments, collections, refunds. Clear the merchant and deposit clearing accounts to zero (guide 04). Reconcile posted cash against deposited cash (guide 06) — this is the step that surfaces underpayment, and the step most often skipped.
Day three — accruals, prepaids and intercompany
Roll the prepaid and accrued schedules forward rather than re-deriving them. Set the expense accruals from the AP cut-off list, not from memory. Then run the intercompany matrix and eliminate (guide 03). Fixed assets and depreciation post here; so do allocations, and allocations post after the pools they allocate are final.
Day four — review the balance sheet, then the P&L
In that order. Every balance sheet line against its support schedule; anything without support is an open item, named. Only then look at the P&L, because a P&L reviewed before the balance sheet is tied is a P&L you will review twice. Flux the P&L against prior month, prior year and budget, and write the explanation while you have the detail in front of you.
Day five — consolidate, report, lock
Consolidated and by-location statements, the variance narrative attached rather than bolted on, and the period closed in the system so nobody posts into it next week. Open items carry forward on a single list with an owner and a date — not in somebody's head.
Leaving the period open "in case something comes in." An open period means the statements you distributed on day five are not the statements in the system on day twelve, and nobody will know which version a decision was made from.
What changes with multiple entities
The chain is the same per entity; what changes is that you cannot consolidate until the slowest entity finishes. So the sequencing question becomes a staffing question: entities run in parallel through day three, and converge on day four.
| Day | Per entity | Group level |
|---|---|---|
| 1 | Cut off, reconcile cash | — |
| 2 | Revenue, AR, clearing to zero | — |
| 3 | Accruals, prepaids, fixed assets | Intercompany matrix, eliminations |
| 4 | Balance sheet review, P&L flux | Allocations, consolidation tie-out |
| 5 | Entity statements, period lock | Consolidated package, narrative |
Intercompany is the single most common reason a multi-entity close slips, and it slips for a structural reason: it cannot be fixed inside one entity's close. Both sides have to be open at the same time, which is why it belongs on day three in every entity rather than whenever each one gets there.
Why closes slip, honestly
No enforced cut-off
If invoices keep arriving through day eight and get posted to the prior month, the close never ends — it just gets reopened. A cut-off that moves is not a cut-off.
Reconciliations done at the end
Reconciling after the P&L has been reviewed inverts the chain. Every correction then invalidates a review that already happened.
One person holds the knowledge
If the close only works when a specific person is available, it is not a five-day close — it is a five-day close plus their calendar. Document it to the point where someone else can run it.
Analysis mixed into production
"While I'm in here, let me look into that variance" is how a two-hour task becomes a day. Note it, finish the step, come back on day four.
The calendar you publish
Write it down, put dates on it, and send it to everyone whose work feeds it — including the people outside finance. Half of a close calendar's value is that the practice managers, the billing lead and the AP clerk can all see the date their part is due, so you are not chasing.
- Dated for the next twelve months, not described in general terms
- Every task has one owner, not a department
- Prerequisites stated, so a slip is visibly a slip in something upstream
- Distributed outside finance, to everyone who feeds a cut-off
- Reviewed once a quarter against what actually happened
The standard I work to. Five business days, consolidated, with the variance narrative attached. Not because faster is automatically better — but because a close that lands on day five is a close whose numbers are still useful for a decision, and a close that lands on day twenty-two is history.