Definition
Matched-record share is the percentage of eligible submitted records for which a provider returns an identified result. Divide matched records by eligible submitted records and keep the provider's status definitions beside the calculation.
The formula is simple: matched records divided by records submitted, times 100. The arguments start over what counts as a match. A vendor that "matched" a record but returned nothing you can use has padded the number without helping you.
matched share shows up everywhere data changes hands. Enrichment projects, CRM appends, ad platform custom audiences, and data appends all get judged by it, and it's usually the first number a vendor quotes in a sales call.
What's a Good matched share for B2B Data?
Measure a single provider and a multi-provider waterfall on the same dated input. Report the denominator, key used, newly matched rows at each step, conflicts, and the records rejected in review.
The key you match on moves the number more than the vendor does.
| Match key | Typical matched share |
|---|---|
| Corporate email address | Measure against a representative known-good sample |
| Website domain (company records) | Usually the strongest company-level identifier; test redirects and parent domains |
| Name + company (no email) | Requires fuzzy matching and manual review of ambiguous candidates |
| Personal email (gmail, yahoo) | Low for B2B enrichment, higher on ad platforms |
| Phone number only | Weakest single key, switchboard numbers repeat |
An ad platform and an enrichment provider may use different reference populations and identifiers. Compare results from the same dated input file, record each denominator and eligibility rule, and investigate which rows matched instead of treating two headline rates as equivalent.
matched share vs. Fill Rate vs. correctness
These three get conflated in vendor conversations, and the conflation is rarely accidental.
matched share tells you how many submitted records the reference process recognized. Fill rate tells you how many requested fields came back populated. Correctness review tests whether the returned values agree with the approved evidence. A record can match while important fields remain blank or a returned value fails review. Report three measures with their denominators, sample method, field definitions, and exception treatment; a matched share alone does not describe usable output.
Build the measurement table at record and field level. Keep the input identifier, matched entity, match key, raw result, review status, and rejection reason. Count ambiguous rows separately rather than treating them as matched or unmatched. This lets a reviewer reproduce the headline rate and see whether one source, segment, or matching rule is driving the result.
When comparing runs, hold the input cohort and acceptance rules constant. If either changes, label the result as a new test rather than a trend. Preserve the dated files so changes can be traced to the records involved.
Loose matching logic inflates the first number at the expense of the third. If a provider counts "someone named J. Smith at a company with a similar name" as a match, the matched share climbs and the error rate climbs with it. When a quoted rate looks too good for the quality of your input file, that's usually what happened.
What Drives matched share Up or Down?
Input quality first. Normalizing before submission is the cheapest matched share improvement available, which is why it's step one in our enrichment process .
Then coverage. Provider databases vary by geography, industry, company size, and field. A low result can reflect the input file, the target segment, or the provider, so test a representative sample and inspect unmatched rows before committing.
Record age matters too. Measure change against a dated baseline instead of applying a general decay benchmark, then set the review cadence from the changes observed in your own system. A low matched share on old records are the decay talking, and it's worth reading our glossary entry on data decay alongside this one.
How to Evaluate a Vendor's matched share Claim
- Ask what key the match used. An exact corporate-email match and a fuzzy company-name match have different ambiguity. Review representative accepted and rejected rows before comparing headline rates.
- Ask what "matched" needed. Recognized the record, or returned usable fields? Get the fill rate for the specific fields you care about, in writing.
- Run a sample. Send 500 representative rows before signing anything. Spot-check 50 of the returned records by hand against LinkedIn and company websites.
- Check a representative set of unmatched records. If reviewers can identify them from the submitted fields, investigate provider coverage; if the rows lack usable identifiers, fix the input or classify them as ineligible.
Example
A useful provider test compares matched, correct, usable, unresolved, and rejected rows on the same input. A provider with fewer raw matches can still return more accepted records. If a waterfall is tested, report each stage separately so the combined result does not hide duplicate or conflicting matches.
What should accompany the matched share?
A matched share is incomplete without the denominator, eligible-record rule, source date, and correctness review. Deliver those fields beside the result. Keep exact, normalized, probabilistic, and manually approved matches in separate status values so the review path remains visible. Examine a labeled sample from each status, including non-matches and ambiguous pairs. Record false positives and false negatives separately, along with the reviewer decision and reason. This prevents a permissive threshold from looking better simply because it accepted more weak candidates.
Frequently Asked Questions
What is matched share?
Matched-record share is the percentage of eligible submitted records for which a provider returns an identified result. Divide matched records by eligible submitted records and retain the status definitions.
How is matched share calculated?
Matched records divided by total records submitted, multiplied by 100. The subtlety is in what counts as "matched." Some vendors count any record they recognized, even if they returned no new fields. Ask whether the denominator is your full file and whether the numerator needed usable data back.
What is a good matched share for B2B data enrichment?
A matched share is meaningful only for a defined input cohort, key, provider configuration, and review standard. Test providers on the same dated sample, report company and person matching separately, and preserve unmatched and ambiguous rows so differences can be investigated.
What is the difference between matched share and fill rate?
Matched-record share measures how many eligible records the provider recognized. Fill rate measures how many requested fields came back populated. Report both by field and segment because a single completion percentage can hide sparse returned records.
Why do ad platform match rates differ from enrichment match rates?
An ad platform and an enrichment provider may use different reference populations and identifiers. Compare results from the same dated input file, record each denominator and eligibility rule, and investigate which rows matched instead of treating two headline rates as equivalent.
How does Verum improve match rates?
Records that fail automated matching get human review rather than getting dropped.
Related Terms
Related: All Glossary Terms | Enrichment Services | Best Data Enrichment Tools