The Close Nobody Owns: Multi-Entity Reporting in Mid-Market Finance

A company with one legal entity has a close. A company with nine has a negotiation.

Entities arrive one at a time — a subsidiary for the German customers, a holding company the lawyers wanted, an acquisition with its own ledger. Each addition looks marginal. Then a controller who used to close in five days is closing in nineteen, and no single decision explains why.

The work is in the seams

The individual entity closes are usually fine. What breaks is everything between them: intercompany balances that don't agree, local statutory treatment that diverges from group policy, translation redone at the wrong rate, a consolidation spreadsheet with a tab named DO NOT DELETE.

None of it appears on an org chart. It has no owner, because it isn't any one entity's job — it's the friction of having several. So it lands on whoever notices first, usually the group controller, at the end of an already long month.

Why headcount is the wrong instrument

The work is a wall of it for eight days a month and comparatively little the rest. Hire for the peak and you pay senior salary for trough capacity; hire for the average and the peak still breaks. Which is how a controller with a decade of technical judgement spends Tuesday chasing an intercompany mismatch of four hundred euros.

The split that works

Reconciliation, tie-outs, currency conversion, mapping local charts to the group chart, chasing the entity that hasn't submitted, assembling the pack — high volume, defined rules, same shape every month.

Then the judgement: whether a local treatment is defensible, how an acquisition's opening balances should be presented, what the auditor will ask and whether the answer is ready. That is why the controller exists.

An AI Specialist on the first, a credentialed accountant owning the second. Not because the software is clever — because the expensive person stops doing the cheap work in the one week their judgement is scarce.

The test

Mark every hour of last month's close as reconciliation or judgement. Most groups find the ratio worse than they assumed, and that the judgement hours were the ones squeezed — done fast, late, and last, because the mechanical work had to happen first.

That inversion is the real cost of a multi-entity structure. It's also fixable without changing the structure at all.