A private equity data warehouse centralizes data. It does not define it.
By Derwin Lucas ·
Summary
A private equity data warehouse brings records from acquired companies into one place, on one schema, where they can be queried together. That is a genuine advance over monthly spreadsheets, and it is the foundation on which comparable reporting can be built. It does not by itself make the data comparable. Mapping each company’s fields to a common column gives them a common name, not a common meaning. The meaning is set by decisions about what counts, when it counts and what is excluded, and those are finance decisions that a warehouse records only if someone makes them.
What a warehouse makes possible
A warehouse holding record-level data from each company changes what a finance team can do. Instead of accepting a submitted close rate, it can count contracts and demonstrations itself. Instead of accepting a reported margin, it can classify costs on one basis. Restating a company’s results under a portfolio definition becomes a query rather than a request.
This is the strongest argument for centralizing data, and it is the reason this paper does not argue against it. Comparability across acquired companies is far more achievable from records than from results.
Mapping fields is not mapping meaning
Warehouse projects typically map each source system’s fields to a common schema. The mapping is technical and usually correct: the field each system uses for a sale date is loaded into one sale-date column.
Exhibit 1. One column, three meanings (illustrative)
| Company | Source field mapped to "sale date" | What the date records |
|---|---|---|
| Company A | Contract date | The day the homeowner signed |
| Company B | Sold date | The day financing was approved |
| Company C | Booked date | The day the rescission period ended |
Each mapping is faithful to its source. The resulting column is consistent in name and format, and a query counting sales by month will run without error. It will also count three different events, assigned to months on three different rules. A contract signed in the last week of a month can appear in that month at one company and the next at another.
The same holds across the schema: a cancellation, an appointment, a completed job and a cost category can each be mapped faithfully and still mean different things from company to company.
Who owns the meaning
The mapping in Exhibit 1 is typically done by whoever builds the warehouse. The question of which event should count as a sale is not a data-engineering question. It determines close rates, revenue timing and incentive pay, and it belongs to the finance leadership accountable for those figures.
When that decision is not made explicitly, it is made implicitly, by the mapping. The warehouse then holds a definition that no one approved, and that changes whenever a mapping is edited.
Private investment data management: what has to sit beside the data
For a warehouse to support comparison, four things need to be recorded alongside the records themselves:
- The portfolio definition of each consequential measure, approved and versioned, with an effective date.
- Each company’s departure from that definition, and whether its records allow restatement.
- The coverage of each figure: which companies, periods and records it includes, and which it excludes.
- The exceptions: records that cannot be classified, submissions that did not arrive, and adjustments under review, stated rather than absorbed.
None of these is data in the ordinary sense. Each is a decision, with an owner and a date. A warehouse that holds the records without the decisions holds the material for comparison, not the comparison.
The financial records
The same gap applies to financial data. A warehouse holding each company’s general ledger mapped to a portfolio chart of accounts has solved the structure. Whether shared-service costs are allocated, where commissions and financing fees are classified, and which adjustments qualify are separate decisions. In an illustrative 25-company portfolio, eight companies that each leave about $150,000 of shared costs out of their EBITDA report $1.2 million more EBITDA between them than they would on the portfolio’s basis, whether their ledgers sit in a warehouse or not. EBITDA comparability across portfolio companies →
When a mapping changes
Mappings are maintained like any other code. A source system is upgraded, a field is renamed, a new company is onboarded with a different configuration, and a mapping is edited to match. Each edit is a definition change for every figure that depends on the column, and unless the warehouse keeps the prior mapping and its dates, the change applies to history as well as to the months that follow.
A sale-date mapping moved from the contract date to the financing-approval date shifts sales between months for every period the warehouse holds. Close rates, revenue timing and year-over-year growth all move, and no submission changed. The control is the same as for any definition: a version, an effective date, an owner, and a record of which periods were rebuilt.
A test that reveals an unagreed definition
A reconciliation exposes definitional gaps directly. For each company and month, the figure the warehouse calculates on the portfolio definition is set beside the figure the company reported on its own. Where they agree, the definitions agree or the difference is immaterial. Where they differ, the difference measures how far the company’s definition sits from the portfolio’s, and the records behind it show why. Run monthly, the same reconciliation shows when a company’s practice changes.
What this paper does not claim
- That a warehouse is unnecessary. Record-level data in one place is a strong starting point for comparability.
- That mappings are wrong. A mapping can be entirely faithful to its source. The issue is that faithfulness to the source is not the same as agreement on a definition.
- That every field needs a definition. The decisions above are needed for the measures behind consequential decisions, not for every column in the schema.
Implication
A data warehouse answers the question of where portfolio data lives. The question of what it means is answered by decisions that sit beside it: definitions approved by finance, departures recorded, coverage stated and exceptions visible. A platform that has built the warehouse has done the harder engineering. What remains is mostly a matter of decision and record.
In Goverisyn, the definition each company applies is recorded, figures are reconciled to their source records, and every figure carries an evidence state: Substantiated, Qualified, Limited or Not Available. What each state means → Goverisyn works alongside a warehouse rather than replacing it. How this differs from a data warehouse →
Goverisyn’s Portfolio Measurement Review tests these decisions on the measures leadership relies on most. The first step is a free 60-minute Portfolio Review, which needs no data in advance.
Start with a Portfolio Review.
60 minutes. No data needed in advance.
By Derwin Lucas, founder of Goverisyn, with nearly three decades in management and information systems in home services. Published . About Goverisyn →