The Coverage Number Is Never the Full Story | NPLUS Global

Every data vendor pitch eventually arrives at a number: 85% coverage, 92% coverage, sometimes higher. It's a useful shorthand, and buyers under time pressure understandably reach for it as a way to compare vendors quickly. The problem is that "coverage" isn't a standardized metric in healthcare or industrial data — there's no regulatory body defining what it measures against, so vendors are free to pick whatever denominator makes them look best. These myths persist not because buyers are careless, but because the sales conversation rewards a clean number over a messy explanation, and messy explanations are usually where the truth lives.

Myth: A coverage percentage tells you how much of your target market the vendor can reach.

Fact: Most coverage claims are measured against the vendor's own database or taxonomy, not against the actual universe of practicing providers, facilities, or organizations in a given segment. A vendor can legitimately claim 90% coverage of "the file they built" while still missing a meaningful slice of, say, independent behavioral health practices or rural imaging centers that never made it into their source pipeline in the first place. Always ask what the denominator is — coverage of what, exactly — before the number means anything.

Myth: An aggregate coverage number applies evenly across specialties, geographies, and org sizes.

Fact: Coverage is almost never distributed evenly; it clusters around whatever the vendor's original data-collection method was built for. A file sourced heavily from claims data or licensing boards will skew toward larger, more institutionally visible organizations and thin out fast once you get into small practices, newer facilities, or specialty niches. If your campaign targets a specific segment, the aggregate number is close to irrelevant — ask for coverage broken out by the segment you actually care about.

Myth: Coverage means the record has usable contact information.

Fact: In most vendors' internal definitions, "coverage" simply means an entity exists in the database — a name, an NPI, an address. It says nothing about whether there's a verified email, a direct dial, or a current decision-maker attached to that record. This is the gap that quietly kills campaigns: the organization is "covered," but the actual person you need to reach isn't represented at all, or is represented by someone who left two roles ago.

Myth: If two vendors both claim 90%+ coverage, their files are roughly comparable.

Fact: Two vendors can both be technically accurate and still be describing completely different realities, because they're measuring against different baselines — one against CMS's NPPES registry, another against a proprietary universe built from web scraping and licensing data, another against whatever their last client contract happened to require. This is one of the more useful questions to ask directly in a vendor evaluation: what's the reference universe, and can you show your math. At NPLUS Global, this is usually the first thing we walk through with a prospective client, because a coverage number without a disclosed baseline isn't really a number — it's a marketing choice.

Myth: Coverage, once measured, stays roughly accurate for the life of the contract.

Fact: Healthcare and industrial organizations churn constantly — practices merge, providers change affiliations, facilities close or rebrand, and org charts shift in ways that don't get reported anywhere centrally or quickly. A coverage figure is a snapshot, and the half-life of that snapshot is shorter than most vendors advertise, especially for fields tied to individual roles rather than static entities like facility addresses. A number that was true at the start of a contract can be materially stale by month six without anyone on either side noticing.

Myth: A vendor's sample file or spot-check proves the coverage claim holds up.

Fact: Samples are, by nature, curated — vendors have every incentive to hand over a sample that represents their strongest segment, not their weakest one. A spot-check of fifty records from a well-covered metro area tells you almost nothing about how the file performs in a thinner region or a smaller specialty. If you want a real read, ask to test against a segment you already suspect is hard to cover, not one the vendor picks for you.

Myth: Combining data from more sources automatically means more coverage, and more coverage means better data.

Fact: Aggregating multiple sources can inflate the count of "covered" entities without actually improving accuracy, because merging isn't the same as reconciling. When three source feeds disagree on an address or an affiliation, the aggregate record often just picks one, silently, and the resulting file looks more complete than it is. More sources can genuinely help, but only if there's real deduplication and conflict resolution behind the scenes — otherwise you're trading a smaller number of honest gaps for a larger number of hidden ones, which is a worse trade, not a better one.