Location encoded in the account number
The classic workaround — 6100-NY and 6100-TX rather than one account with a location dimension. It works until you open a fourth site and have to add forty accounts.
By structure
Multi-location businesses ask for the same thing almost universally: profitability by site, monthly, without somebody assembling it. Whether that is easy or hard turns on one question nobody asks early enough — is a location a dimension or a legal entity? Those have very different answers.
The situation
The classic workaround — 6100-NY and 6100-TX rather than one account with a location dimension. It works until you open a fourth site and have to add forty accounts.
The largest cost at most locations enters the ledger as one summarised journal with no location on it, which makes site P&L impossible no matter how well everything else is coded.
Head office, insurance, software, and marketing consumed by every site with no agreed basis for splitting them, so site margin is arguable and therefore ignored.
The people who could act on site performance get a monthly PDF three weeks late, or nothing, because permissions were never scoped by location.
Frequently for insurance, licensing, or partnership reasons. Mixing dimensions and entities in one reporting model is where most of the complexity actually lives.
Where sites keep their own books, the group view arrives when the last one finishes, and nobody can see which one that is.
A location that is a dimension is straightforward. It is a field on every transaction, and location P&L becomes a query. Adding a site is adding a value to a list.
A location that is a separate legal entity is a different problem. It has its own books, its own close, intercompany activity with the others, and possibly its own currency. That needs consolidation with elimination rather than a filter.
Most multi-location businesses have some of each, and the mistake is treating them uniformly — either forcing entities into a dimension model, which loses the elimination, or making every site an entity, which multiplies closes for no benefit.
Labour is the largest cost at most sites, and it typically enters the ledger as one summarised journal from the payroll provider with no location on it. That single gap makes site P&L impossible regardless of how carefully everything else is coded.
Mapping payroll to location at the employee level is a mapping exercise rather than a technical one, and it is frequently the highest-value hour in a multi-location engagement. We check for it during the assessment because discovering it in week four wastes the preceding three.
Head office, insurance, software, and marketing are consumed by every site and the basis for splitting them is a management decision — headcount, revenue, square footage, or transaction volume. There is no correct answer and the failure is not choosing.
Where no basis is agreed, site margin is arguable and therefore ignored by the people it was produced for. We push for an agreed method early, apply it as real journal entries rather than a reporting adjustment so the numbers tie, and show the method on every report.
The most underused improvement is scoping permissions by location so a site manager sees their own P&L, live, rather than receiving a PDF late. It changes site performance more than any reporting refinement, because the person who can act on it is finally looking at it.
Questions
Tell us how many sites and which are separate entities, and we will scope what it actually takes.