The first working day of the month. Someone in finance exports the sales orders from the ERP, pastes them into a spreadsheet, adds the hours from the time registration system and emails the whole thing to a colleague for checking. Meanwhile the close is waiting for the last supplier invoices, for a correction from last month and for the one colleague who knows how the intercompany entries work. A week later the figures are ready. Sometimes two.
Month-end close rarely takes long because of the closing itself. It takes long because of what did not happen in the month before: figures that never reached the ledger but ended up in exports, spreadsheets and inboxes. At the end of the month, collecting and reconciling becomes the work, when it should have been checking.
Where the days go: collecting, reconciling and waiting
Ask a controller where the time goes and the answer is almost always the same. First the collecting: bringing data from the ERP, the CRM, time registration and the bank together in one file. Then the reconciling: checking that the revenue in the CRM matches what was invoiced, going through open items, netting intercompany transactions. And in between, the waiting: for the department that has not submitted its hours yet, for the supplier who invoices late, for the approval of a manager who is in a meeting.
Of the three, waiting is the invisible one. It appears on no timesheet, yet it determines the lead time more than the work itself. A close that takes a week or two rarely contains that many days of work; most of it is waiting for each other.
The same corrections every month
The second part of the delay is the corrections. Every manual transfer, from an export to a spreadsheet, from a spreadsheet to the accounting package, is a moment where something can go wrong: a sales order that was not carried over, an amount with a misplaced decimal point, a version of the file that turns out not to be the latest. Those errors surface when the figures do not reconcile, and then the searching begins. How that mechanism works is explained in every manual data transfer is a chance of an error.
On top of that, much of this work lives in people’s heads. One colleague knows which correction is needed every month and why; it is written down nowhere. When that colleague is ill or on holiday, the close stalls or someone else does it slightly differently. That makes month-end close not only a process issue but also a dependency on a few people.
The close is only as fast as the month before it
Organisations with a short close do not work harder in the first week of the month. They have moved the work. Sales orders are already in the ledger the moment they are created, supplier invoices are booked as soon as they are approved, hours flow through to invoicing without anyone retyping them. At the end of the month there is nothing left to collect. What remains is checking and judgement: reviewing deviations, assessing provisions, explaining the report.
In finance this is called a continuous close: the figures are largely right throughout the month, and the close is a formal confirmation rather than a reconstruction. You do not need the term to apply the principle. The only question is which flows are manual today and which of those can be automated.
What can flow automatically during the month
Most of the delay comes from a handful of flows that are processed by hand every month:
- Sales orders and invoices from the ERP or CRM into the ledger, so revenue is booked the moment it arises.
- Supplier invoices booked directly after approval, including the link to the right cost centre or project.
- Hours from time registration into invoicing and project administration, without a detour through a spreadsheet.
- Bank transactions matched automatically against open items, so chasing payments is no longer a day’s work.
Common accounting packages offer integration options for automating flows like these, and so do most ERP, CRM and time registration tools. The integration itself is then the execution; the thinking step before it is deciding which system is leading for which data, so the ledger is not fed from two sides at once.
A dashboard does not speed up the close
The reflex is often a BI dashboard: once the figures are on one screen, the rest will follow. But a dashboard shows what is in the source systems. If the figures are still in spreadsheets and inboxes, the dashboard shows an empty or outdated picture, and the close takes just as long. The dashboard only becomes useful once the flows beneath it are in order. What happens next, and why management otherwise tends to steer on old figures, is covered in the board decides faster than the reporting can keep up.
Want to know what your close is waiting for? For one month, record per step who is waiting for whom and which data is transferred by hand. That overview usually points to two or three flows that determine the lead time. Those are the flows we connect directly to the ledger with API integrations, so the next close starts with figures that are already there.