Written by: Karol Sobieraj, Founder & CEO, Digital Colliers
If your month-end close runs eight days, you're not slow. You're normal. FSN's research puts the mid-market at 8-10 days, and that's been true for a decade. APQC's benchmarks put the top teams under five. So the gap is real, it's roughly half, and it's not closing on its own.
The interesting question isn't why you take eight days. It's why the gap between you and the fast teams has barely moved since 2015, despite every ERP vendor promising a faster close.
The four-system reality
Walk into any mid-market finance function and you'll find the same picture. The GL sits in one system. Billing sits in another. The CRM has ARR and contract data that doesn't quite match billing. Payroll and expenses live somewhere else again. Nobody wrote it this way on purpose. It grew.
FSN's work on this is pretty blunt: most mid-market teams still run close in spreadsheets, pulling numbers across systems by hand. That's the actual workflow. Someone exports from NetSuite. Someone else pulls a Salesforce report. A third person reconciles against the bank feed. Then a controller stitches it together in Excel and chases the variances.
The eight days aren't spent closing books. They're spent making four systems agree on what happened.
Why senior hires don't fix it
This is the part CFOs learn the expensive way. You hire a strong controller from a faster shop. You give them six months. Close still takes eight days.
The reason is structural. A senior finance hire can tighten the process, kill some manual journals, and push the team harder. What they can't do is rewrite the plumbing between your billing system and your GL. That's an engineering job, and finance leaders rarely have engineers reporting to them.
So the new hire optimises the visible workflow, which is the last twenty percent. The invisible eighty percent, the data movement, stays exactly where it was.
This is also why so many finance automation projects stall. Roughly 95% of enterprise AI and automation projects don't reach production, according to MIT's recent work. The ones that die tend to die at the integration layer, not the model layer. Somebody built a nice dashboard on top of data that still requires three humans to assemble.
The plumbing question test
Here's a quick diagnostic. Ask your finance team three questions:
When revenue in the CRM doesn't match revenue in the GL at month-end, who reconciles it, and how long does it take?
If a contract is amended mid-month, how many systems need to be updated by hand?
When the auditors ask for a transaction trail, does someone run SQL by hand, or does a report exist?
If the answers involve names of people rather than names of systems, you have a plumbing problem. No amount of process discipline will get you under five days. You're solving the wrong layer.
What actually changes month-end
The teams that close in four or five days did one specific thing. They stopped treating the gap between systems as a finance problem and started treating it as a data problem.
In practice that means a few things:
One source of truth for each entity. Customers, contracts, transactions, and journal lines each have exactly one system that owns them, and the others read from it.
Automated reconciliation that runs daily, not monthly. If billing and GL are out of sync, you find out on day two, not day thirty-two.
Exception queues instead of full manual review. Humans look at the 3% that don't match, not the 97% that do.
Audit trails as a byproduct of the pipeline, not a separate scramble in week two.
None of this requires ripping out your ERP. It requires someone who can read your data model, write against your APIs, and build the connective tissue between what you already own.
The cost of doing nothing
The cost isn't the four extra days. It's what those four days hide. You can't run a weekly forecast if your last actuals are twelve days old. You can't answer a board question about margin by segment without a two-day pull. You can't pass a DORA operational resilience review with a close that depends on three people and a spreadsheet, and DORA has been in force since January 2025.
Eight days of close is a symptom. The disease is that your finance data doesn't move on its own. Fix that, and the close fixes itself.
Sources
This article was originally published on the Digital Colliers Blog. Digital Colliers helps DACH and UK companies implement AI — see our AI consulting services or contact us.
Top comments (0)