Bramvia — Business Central Experts
Why does our month-end close take three weeks? The seven causes, in the order to fix them
A close beyond ten days is a recognised signal of data and process problems in finance, purchasing and inventory. The seven causes we find, what each one costs, and why the fix is usually configuration rather than a new system.
Short answer: a financial close that runs past ten days is treated in diligence and in ERP practice as a reliable signal of data and process inefficiency across finance, purchasing and inventory — not as a finance-team problem. The causes are almost always the same seven, and they sit upstream of accounting: goods received but not invoiced, inventory nobody trusts, intercompany that doesn't agree, accruals estimated by hand, bank reconciliation done in a spreadsheet, a reporting pack assembled rather than produced, and approvals that happen by email. Fixing them in the right order usually takes weeks of configuration rather than a new system. And the payoff is bigger than time: management decides on July's margin in early August instead of late September, and a buyer or lender reads a five-day close as a finance function that scales.
What a slow close actually costs
- Decisions made a month late. Pricing, purchasing, credit decisions taken on stale information. This is the real cost and it never appears as a line.
- The finance team's month. Three weeks of close means the team has one week for analysis. You're paying for accountants and getting bookkeeping.
- Valuation and credit. Diligence and lenders read close speed as a proxy for data quality. Why that matters before an exit.
- Audit cost. Manual adjustments and reconciliations lengthen fieldwork. Surveys put roughly a third of mid-market CFOs still relying on manual processes for audit preparation.
- Errors found late. A mistake discovered on day 19 has already informed three weeks of decisions.
The seven causes, in fix order
1. Received not invoiced. Goods arrived, the invoice hasn't, and nobody knows the value sitting in that gap. Every close starts with a hunt. Fix: three-way matching with receipts posted at the dock, and an accrual driven by the receipt rather than by memory. Configuration, not development — and the payables agents in modern ERPs now do most of the matching.
2. Inventory nobody trusts. If the stock figure is a negotiation between the system and the warehouse's spreadsheet, cost of goods is an estimate and so is your margin. Fix: cycle counting instead of an annual count, directed putaway and picking so movements are recorded when they happen, and bin logic a new hire can follow — which matters when warehouse turnover runs near 49% a year.
3. Intercompany that doesn't agree. Two entities, two versions of the same transaction, one reconciliation nobody enjoys. Fix: intercompany posting in the system with matched documents. If you have two ERPs after an acquisition, this is one of the costs of leaving them separate.
4. Accruals and cut-off by hand. Freight in transit, utilities, commissions, rebates — estimated in a file each month. Fix: recurring journals with templates, and accruals derived from open documents. At current freight rates, in-transit freight is no longer a rounding error.
5. Bank reconciliation in a spreadsheet. Statement exported, matched by hand, differences chased. Fix: bank feeds with automatic matching. This is often a single afternoon of setup that removes two days from every close.
6. The reporting pack assembled, not produced. Exports, pivot tables, formatting, and a version emailed around. Fix: financial reports defined once in the system, and dashboards that refresh themselves. In Business Central the four Power BI apps are included, and version 29 adds report packages delivered as a single scheduled PDF — the cure for "the pack" specifically.
7. Approvals by email. Invoices, credit limits and journals waiting in inboxes, with no record of who approved what. Fix: approval workflows in the system, with the audit trail as a by-product rather than a reconstruction.
Why the order matters
Most companies attack number 6 first, because the pack is the visible pain. It's the wrong end. A faster report over unreliable numbers gets you to the wrong answer sooner. Causes 1 to 5 are where the days and the accuracy live; 6 and 7 are how you stop re-doing the work once the data is right.
There's also a sequencing reality: fixing 1 and 2 changes the numbers. Better to discover that before you've built the reporting on top of them.
How to find out which of the seven you have
Run one close with a stopwatch. Not a project — just record, for each step, who did it, how long it took, and what they were waiting for. Two things come out of it:
- The critical path. Usually two or three steps consume most of the elapsed time, and they're rarely the ones people complain about.
- The waiting. Most of a slow close is not work; it's waiting for information from outside finance. That tells you the fix is upstream.
Bring the list to whoever owns the system. If the answer to most items is "the system can do that, it was never configured," you have weeks of work ahead, not a migration. That's the common case — and it's the cheapest engagement anyone will ever sell you.
When it really is the system
The honest minority. If your ERP has no three-way matching, no intercompany posting, no bank feeds and no reporting layer, no amount of configuration produces a five-day close. Then the question becomes whether to keep paying the monthly cost of the workarounds or to move — a decision that should follow the stopwatch exercise, not precede it.
How we help
The stopwatch exercise is part of our free assessment, and we'll tell you plainly which of the seven you have and which are configuration versus development. If you want the cost quantified — hours consumed, decisions delayed, cash affected — that's the Profit Leak Audit, fixed fee, on any ERP. And when there's build work, each part is invoiced only after you accept it.
FAQ
What's a realistic target? Five business days for a mid-market single entity; seven to ten with several entities and intercompany. Under three usually means either a very clean operation or something not being checked.
Our close is slow because we're understaffed. Sometimes true. But if most of the elapsed time is waiting rather than working, hiring doesn't fix it — and the stopwatch exercise tells you which it is.
Can AI speed up the close? The useful parts already ship in modern ERPs: agents that match vendor invoices to orders and receipts, and validation that flags anomalies before posting. They help with causes 1 and 4. They don't fix inventory nobody counted.
Do we need to change ERP to close in five days? Usually no. In most companies we assess, five of the seven causes are configuration gaps in the system they already own.
Close taking three weeks? Free assessment, no commitment — bring one close's worth of notes and we'll tell you where the days are going.