Why your roll-up has five versions of revenue
Somewhere around the second or third acquisition, every healthcare roll-up has the same meeting. The board asks why a number moved. The room goes quiet. Four days later, there’s an answer, assembled by a controller who rebuilt the same spreadsheet she rebuilt last month, from exports nobody fully trusts.
The instinct is to blame the people or the tools. It’s neither. It’s that every company you bought came with its own systems and its own definitions, and the definitions are the part nobody budgeted for.
The same words mean different things in every entity
Take three metrics any healthcare services operator uses daily: a working provider, a net booked week, time-on-assignment. Does a clinician who worked one shift this month count as working? Does a booked week net of cancellations, or of cancellations and no-shows? Does time-on-assignment start at contract signature or first shift? Each acquired entity answered these questions years ago, independently, inside its own billing system, and every answer is defensible. They’re just not the same answer.
Roll five of those up and you don’t have a revenue number. You have five revenue opinions in a trench coat.
Three failures, three different fixes
What presents as “our reporting is a mess” is usually three distinct failures, and they don’t share a fix:
No single source. Each entity has its own system and its own definitions. The fix is plumbing: one warehouse, one conformed definition of provider, client, encounter, revenue.
No trust. Numbers exist, but leadership keeps its own spreadsheets because the two never agree. The fix is not more dashboards; it’s reconciliation: tying the new numbers line-by-line to the report leadership already relies on, and re-checking on every load.
No speed. Every question costs a person and a week. The fix is self-serve: dashboards and plain-English query on top of governed data, not on top of the mess.
This is why buying a BI tool changes so little. If five systems disagree, five dashboards disagree faster.
What this actually costs
Two examples from inside the work. At one acquisitive workforce platform, permanent-placement fees lived entirely outside the reporting path: booked, banked, and invisible in every board metric. Modeling them in surfaced $559K of revenue in a single line of business that no leadership report had ever counted.
At the same platform, a validation harness (the automated reconciliation described above) caught a silent upstream extract change that had swung two brands’ numbers by +68% and −91%. Nobody had touched the dashboards. The numbers just quietly became wrong, upstream, overnight. Without the harness, the board would have seen both.
Where to start
Not with the platform. With the definitions. Write down every metric leadership uses, how each entity computes it today, and which computation wins going forward, and get it signed. It’s unglamorous, it takes about two weeks, and it’s the step every failed analytics project skipped. The plumbing is real work, but it’s known work once the definitions exist.
If you want to know which of the three failures you actually have, that’s a two-week question, not a two-quarter one.
Kaleb Lewis