Who makes your monthly report?

Partly is the norm and precisely the problem

Early in 2026 Exact asked more than 1,700 owners and managers of Dutch SMEs how far their business processes are automated. For reporting the answer is 26% fully, 61% partly, 13% not at all. Payroll is the only process that runs fully automated at almost half of the companies. Bookkeeping sits in between, with a third fully automated.

In the same survey, one in four names better and more current insight into their own financial figures as one of the biggest challenges in financial administration. After late-paying customers, administrative burden and regulation, it is the first challenge that is about the figures themselves.

Those two numbers belong together. Reporting that is partly automated is as current as the last time somebody built it.

What partly looks like on a Tuesday morning

Partly automated looks the same at most companies. The accounting package delivers an export of the ledger entries. Time tracking delivers a second file, the CRM a third, the web shop or the warehouse a fourth. Someone puts them in Excel, maps the accounts to the layout of the monthly report, corrects what does not reconcile and turns it into a PowerPoint.

That work is not stupid and it is usually done well. It has two consequences nobody wants. The report is ready when the month is already a week behind you. And a question that is not in the standard report costs a week, because the exporting starts over.

That is what the 24% mean by more current insight. They are asking for a report that does not have to wait for someone to build it.

Why it stays at partly

Automation arrives one system at a time. The accounting package automates the bookkeeping and ships a dashboard with it, for liquidity and open invoices. Time tracking has its own overview, the CRM its own pipeline. Each package does that well, for what sits inside it.

The report the board wants combines those systems. Revenue from the books against hours from time tracking, per project, per site, against the pipeline from the CRM. No package can build that from the inside, because the other half of the number lives somewhere else. So a person does it, in Excel.

That is why reporting gets stuck at partly while payroll and bookkeeping move on. Payroll is a closed process inside one system. Reporting runs across all of them. Every system you add makes the manual part bigger, not smaller.

So the 61% is not a lack of effort. It is the shape reporting takes as long as nothing sits between the systems and the report.

The layer that belongs in between

What is missing is a business data model: the layer where all the systems come together and where the meaning of the figures is settled, with the reporting on top. It has four parts.

Extraction pulls the data every night from every system that holds figures, as it is. Data warehouse is the database where that data lands side by side, organised the way your business works rather than the way the packages store it. Definitions & KPIs is where it is settled once what revenue is, which customer is the same customer and how margin is calculated. The KPIs are worked out there too. Reporting is what sits on top: Power BI, Microsoft Fabric or the tool you already have. It only reads.

For the reporting that changes one thing and the rest follows from it: the report no longer fetches anything. It reads from the model. That model was refreshed this morning.

So yesterday's figures are there every morning without anyone having built anything. A question outside the standard report is a new view on the same model, not another round of exports. The number in the sales report is the number in the month-end close, because it is calculated in one place. And what is fixed after the close stays fixed.

That is the difference between 61% and 26%. The report did not get smarter. It just no longer has to fetch anything.

Three objections we hear often

"Our accounting package already has dashboards." It does. They are good for what sits inside the package: liquidity, receivables, the profit and loss at ledger level. The moment a number has to come from two systems, they stop. The package dashboard stays where it is. The model goes underneath, for the reporting the package cannot build.

"That is something for large companies." A business data model costs around €3,935, €8,350 or €19,250 with us, depending on the number of systems and the size of the reporting. With two source systems the first report is live after three to six weeks. That is the scale of a company with twenty to two hundred and fifty people, which is where we prefer to work.

"Excel is staying anyway." Yes. For analysis, checking and a scenario you want to run through quickly, Excel is the right tool. The difference is where Excel gets its figures: from the model, instead of from six exports somebody first had to collect.

Where to start

With the Business Data Scan. One week in which we map your systems and definitions, ending in an action plan with a fixed price for the data model. The scan costs €1,200 and what you keep is yours.

The first question in that week is which reports you currently build by hand every month and from how many systems. The answer usually already shows where the 61% sits in your company.