The Identity Resolution Problem Nobody Budgets For

Every “single view of the customer” initiative I’ve seen starts the same way: with a platform decision. Which CDP, which data warehouse, which reference architecture. The roadmap gets built around the tool, the budget gets allocated to licenses and integration, and the demo looks beautiful — one clean, unified customer profile glowing on a screen.

Then the program meets reality, and reality asks a deceptively simple question: is this the same person?

That question — identity resolution — is where Customer 360 programs actually succeed or fail. And it’s almost always the line item nobody budgeted for.

“Single view of the customer” is an identity problem wearing a data-integration costume

On the surface, unifying customer data sounds like plumbing: pull records from the systems, land them in one place, join them together. The trouble is the join. The same person shows up as “Robert Smith” in one system, “Bob Smith” in another, and “R. Smith” at an old address in a third. A married couple shares an account here and holds separate ones there. A small-business owner is both an individual client and a business entity. Someone moved, changed a name, or was entered twice by two branches on the same day.

There is no universal key that ties all of this together. So the “join” isn’t a join at all — it’s a series of judgment calls about which records represent the same human being, the same household, the same relationship. That’s identity resolution, and it is the actual work.

In financial services, it’s harder — and it matters more

Wealth and banking add layers most industries don’t have. You’re not just resolving individuals; you’re resolving households, advisor-client relationships, beneficial ownership, joint and trust accounts, and the hierarchy between a person, their entities, and their relationships. A “customer” is often a web, not a row.

And the stakes are higher. Get identity wrong and you don’t just send a duplicate email — you misstate a household’s assets, you fragment a relationship an advisor is trying to deepen, you miscount your own customers to the executives who report those numbers, and you feed contradictions into every risk, compliance, and AI system downstream.

Why it’s chronically under-budgeted

The reason is simple: identity resolution is invisible in the demo. The platform vendor shows you a pristine unified profile built on clean sample data. What they don’t show is the 20% of records that don’t match cleanly — the edge cases, the near-duplicates, the ambiguous merges — which is exactly where the cost, the risk, and the months of work actually live.

So the money goes to the platform, because that’s what you can see and buy. The resolution logic — the matching rules, the survivorship logic that decides which value wins when records conflict, the stewardship process for the cases a machine can’t confidently call — gets treated as a detail. It is not a detail. It is the program.

What actually works

The teams that get this right treat identity as a first-class data product, not a feature of the platform:

• Deterministic and probabilistic matching. Exact keys where you have them; scored, fuzzy matching where you don’t — tuned, not accepted out of the box.
• Survivorship rules you can defend. When two records disagree, something has to decide which value becomes the golden record — and you need to be able to explain why.
• Stewardship for the gray zone. Machines resolve the confident cases; people resolve the ambiguous ones. Budget for the humans, not just the software.
• Households and hierarchies as design requirements, not an afterthought bolted on later.
• Measurement. Match rates, and — just as important — false merge rates. Wrongly combining two different people is worse than failing to combine one person, especially under regulatory scrutiny.
• Continuous maintenance. Identity decays. People move, marry, open and close relationships. Resolution is a running capability, not a one-time cleanup.

None of that is glamorous. All of it is what makes the unified profile real instead of a demo.

The part that ties back to everything else

Here’s why this matters more now than it used to: every downstream capability inherits your identity decisions. Marketing activation targets whoever your resolution says exists. Personalization and next-best-action reason over the profile you assembled. And AI models learn from it — which means a model trained on phantom or wrongly merged customers learns confident nonsense. Your Customer 360 is the foundation your AI stands on; if the identity underneath is shaky, so is everything built on top.

The platform is the easy part — it’s the part you can buy. Resolving who’s who is the part you have to actually do, and it’s where the whole promise of a single customer view is won or lost.

Budget for it accordingly. The teams that fund the identity problem as seriously as they fund the platform are the ones whose Customer 360 the business actually trusts.