The Data Quality Problem No One Wants to Own

Walk into almost any finance analytics project that disappointed and you'll find the same autopsy result. The tool was fine. The dashboards were well-designed. The forecasting logic was sound. What killed it was the data underneath—inconsistent, incomplete, defined differently in different systems, and trusted by no one. The analytics layer was built on sand, and no amount of polish on the surface fixes a foundation problem.

Data quality is the unglamorous root cause of more failed finance technology than any other single factor. And the reason it stays unfixed is that nobody wants to own it.

Why data quality becomes an orphan

The problem is structurally homeless. IT owns the systems but not the meaning of the data inside them. The business units own the operations that generate the data but not its consistency across the company. Finance consumes the data but historically hasn't seen producing clean data as its job. So the data quality problem sits in the gap between everyone, owned by no one, getting slowly worse.

Meanwhile, each system develops its own version of the truth. The customer who's "ACME Corp" in one system is "Acme Corporation" in another and "ACME" in a third—now they look like three customers. The chart of accounts means something slightly different in the subsidiary than at the parent. The product hierarchy in the ERP doesn't match the one in the sales system. None of these are catastrophic alone, but together they make any cross-system analysis an exercise in reconciliation and any single number a subject of debate.

The "single source of truth" that doesn't exist

Every company says it wants a single source of truth, and almost none have one—because building one requires resolving exactly the definitional conflicts that nobody owns. A single source of truth isn't a technology you buy. It's a set of decisions: this is how we define a customer, this is our canonical chart of accounts, this is the one product hierarchy, and these definitions are maintained in one place and consumed everywhere.

Those decisions are organizational, not technical. They require someone with authority to say "this is the definition" and make it stick across business units that each have their own preferred version. That's why it's so often left undone—the technical part is tractable, but the governance part requires authority and persistence that's uncomfortable to exercise.

Why the CFO should own it

Here's the argument for finance taking ownership: finance is the function that most depends on data being consistent across the whole company, and the CFO is usually the executive with both the cross-functional view and the authority to enforce definitions. Finance has to roll up data from every part of the business into one coherent picture. That makes finance both the biggest victim of poor data quality and the natural owner of fixing it.

This doesn't mean finance does all the work—it means finance owns the standard. Finance defines what a clean customer record looks like, what the canonical hierarchies are, what the data has to satisfy to be trusted. The actual maintenance can live with the systems and the business units, but the standard and the accountability sit with finance.

Start where the pain is sharpest

You don't fix all data quality everywhere at once—that's a multi-year boil-the-ocean project that loses momentum. You start with the specific data that feeds your most important decisions and is most obviously broken. Usually that's customer master data, the chart of accounts, and a product or cost hierarchy. Get those defined canonically, maintained in one place, and consumed consistently. Prove the value on the high-pain areas, then expand.

The payoff compounds. Once the core data is clean and consistently defined, every analytics tool, every dashboard, every forecast that consumes it gets more reliable at once. You stop reconciling systems and start trusting numbers. The disappointing analytics projects start working—not because you bought better tools, but because you finally fixed the foundation they stand on.

Somebody has to own data quality, and the honest answer is that it should be you. It's the least glamorous thing on the CFO's plate and quietly one of the most valuable.