The starting point

The group: industrial services, 14 legal entities across seven countries, ~€420m revenue, on a Tier-1 ERP since 2022. The CFO had inherited the system in mid-2025. By the time we were brought in, the symptoms were textbook:

  • Monthly close: 22 working days, with consolidation slipping into the next reporting cycle.
  • Four reconciliation tools running outside the ERP — including one Excel-and-VBA model the previous controller had built and never documented.
  • Intercompany imbalances of €2-4m at month-end, manually levelled before consolidation.
  • Audit findings on multibook depreciation and tax provisioning — partially booked at group, partially at entity, with no consistent rule set.
  • Two open vendor proposals: "re-implement on Platform A" and "migrate to Platform B". Combined budget request: north of €6m, 14-month runway.

The board's instinct was to rip and replace. The CFO suspected the system was not the actual problem. She asked us to check that suspicion before committing to either vendor proposal.

What we found in two weeks

The ERP was fit for purpose. The architecture sitting on top of it was not. The platform was being asked to compensate for finance design that had never been done.

The diagnosis

The two-week assessment looked at five layers — operating model, finance architecture, master data, system landscape, governance — and produced a single page with the target architecture and a defendable plan. The headline findings:

1. The chart of accounts had been copied, not designed.

14 entities, 14 different operational reads of the same accounts, no group-level account dictionary. Mapping to consolidation was happening in spreadsheets, every month, by hand.

2. Multibook had never been turned on.

Local GAAP and IFRS adjustments were being booked into the same ledger using "tagging" conventions that drifted from country to country. The ERP supported multibook from day one. No one had configured it.

3. Intercompany was a process, not an architecture.

Every IC posting was a two-sided manual entry. No automated counter-booking, no central IC clearing entity, no policy on which entity led which transaction. The €2-4m imbalance was the residue of that.

4. The "four reconciliation tools" were symptoms.

They existed because the close was broken upstream. They had become entrenched because the people running them were the only ones who understood what was happening to the numbers.

5. Governance had no owner.

No release calendar, no change advisory, no documented finance ownership of the ERP. Every adjustment was a one-off conversation between controlling and IT.

The board was being asked to fund a €6m system replacement to fix problems that were architectural, not technical. The platform was carrying blame for decisions that had simply never been taken.

What we changed

We proposed — and the board approved — a finance-architecture programme on the existing platform. Three workstreams, six months, fixed-fee assessment block plus phased delivery.

Workstream A — Finance design

  • Group account dictionary, 380 accounts, mapped 1:1 to consolidation.
  • Multibook activated: local GAAP + IFRS as parallel ledgers, single source.
  • Tax architecture redesigned per jurisdiction, with clear local-vs-group rules.
  • Consolidation logic moved from spreadsheet to native ERP consolidation.

Workstream B — Intercompany

  • Central IC clearing entity introduced.
  • Automated counter-booking for AR/AP, intercompany invoices, recharges.
  • Documented IC policy: who leads, who follows, how disputes are resolved.

Workstream C — Governance and operations

  • Finance-owned release calendar, monthly cadence, change advisory board.
  • Three of the four reconciliation tools retired. The fourth — for treasury — was kept and properly integrated.
  • BPO support stood up for two of the smaller entities, freeing local controllers for higher-value work.

The outcome, six months later

  • Monthly close: 22 days → 8 days. Group consolidation back inside the reporting cycle.
  • Intercompany imbalance at month-end: €2-4m → under €100k, almost entirely timing-driven.
  • Audit findings on multibook and tax provisioning closed.
  • Three reconciliation tools decommissioned; one remaining tool properly integrated.
  • Total programme cost: well below the €6m re-implementation proposal — and the platform stayed.

Engagement principle

When a finance close is broken, the platform is rarely the right thing to replace. Replace the architecture sitting on top of it. The platform — almost always — survives.

Why this matters beyond this engagement

The pattern repeats. A multi-entity group, four to six years into an ERP, running an increasingly broken close, surrounded by vendor proposals to re-implement. In a meaningful share of those situations, the ERP is fine. The finance architecture was never finished. Re-implementing produces the same problem on a different system, two years from now.

Inside the group, this is the kind of work OPCON and BPO carry — execution and finance operations, plugged into whichever ERP the customer is on. The platform-specific depth (NetSuite, Microsoft, SAP, Zoho, Xentral) sits in the corresponding business units. The diagnosis stays the same on every platform: architecture first.