State of B2B data quality 2026: a verification-status benchmark

The sample selected completed rows where queue_id modulo 101 equaled zero from the production verification queue. That produced 26,988 unique email-status pairs. We joined them to the current contact index by normalized work email; 21,641, or 80.2 percent, matched a current record with usable industry evidence.

Industries were grouped by documented keyword rules into broad reporting categories. A record can contain more than one source industry; the first matching reporting rule assigns the row to one category so denominators remain additive.

A verifier outcome is measured before send. A bounce is an event returned by a recipient system after send, and it can include policy, reputation, throttling, content, or infrastructure causes beyond mailbox existence. Converting the invalid column directly into a predicted bounce rate would be false precision.

Financial services and professional-services samples show especially high unknown shares. That may reflect receiving-domain policy, verifier visibility, or source composition. It does not prove that those mailboxes are dead.

Use the table to size verification workload and to detect status-mix drift. Re-run the same query every quarter, publish the denominator, and compare like with like. For a provider benchmark, hold the input list and verifier set constant.

Historical operator-supplied tests of Apollo and ZoomInfo exports are discussed separately because they used vendor-specific samples and multiple verifiers. They are not mixed into this production-queue industry table.