Key finding
Of 655,442 B2B email addresses put through verification, 54.0% came back verified as deliverable, 18.1% resolved to catch-all domains, 14.6% bounced, 9.4% could not be determined, and 4.0% were pattern-generated guesses.
What we measured
"Has an email" and "can be emailed" are different claims, and the entire list-selling industry lives in the gap between them. Deliverability statistics exist, but they are published by email-verification vendors selling verification, on samples nobody can inspect.
This is the same measurement from the other side of the table: a B2B data operator publishing the verification outcomes of its own database's email addresses, gap included. Every address in the sample was run through standard verification (mailbox and domain checks); this study reports what came back.
Methodology
Denominators, stated in order. The database holds 2,037,030 contacts (audited in full here). Of them, 1,070,952 (52.57%) have an email address on file. Of those, 655,442 (61.20%) have been through verification; addresses enter verification as they are used or enriched, so the checked subset skews toward records that campaigns actually touched. Every status share below divides by the 655,442 checked, and no figure in this study should be quoted against a different denominator.
Status definitions. Verified: the mailbox accepted the check. Catch-all: the domain accepts any address, so the specific mailbox cannot be confirmed or denied. Bounced: the check failed hard. Unknown: the check could not complete. Pattern-generated: the address was constructed from a name-format pattern rather than observed, and has not been confirmed.
A data note, disclosed rather than cleaned. 1,414 records carry a verification status but no stored email (the address was removed at some point after its check). They are 0.2% of the checked set and are left in the counts.
The findings
| Verification outcome | Addresses | Share of checked |
|---|---|---|
| Verified deliverable | 353,834 | 53.98% |
| Catch-all domain | 118,308 | 18.05% |
| Bounced | 95,700 | 14.60% |
| Unknown | 61,296 | 9.35% |
| Pattern-generated | 26,304 | 4.01% |
Two derived figures, labeled as derived: verified addresses are 33.04% of all stored emails, and 17.37% of all contacts in the database. The resemblance of that last number to the database's 17.28% decision-maker share is a coincidence of this snapshot, not a relationship.
The hard-bounce rate deserves its own sentence, because it is the number teams feel first: roughly one in every seven checked addresses (14.60%) in a professionally maintained B2B database failed outright.
What this means
The stack of denominators is the finding. Start with 100 B2B contacts: 53 have an email at all, roughly 28 of those emails would verify as deliverable if checked at the observed rate (52.57 x 53.98%, derived), and the rest scatter across catch-all, bounce, and unknown. A "two-million-contact database" is, for cold email purposes, closer to a 354-thousand-confirmed-deliverable database, and every vendor's numbers work this way whether or not they publish the stack.
The catch-all fifth is the strategically annoying part. Those 118,308 addresses are not bad; they are unknowable in advance, and teams must choose between skipping a fifth of their addressable market and accepting bounce risk inside it. No verification product resolves this; it is a property of how those domains are configured.
B2B email decay is the standard explanation for the bounce rate, and it fits: people change jobs, mailboxes die, and a database's addresses age continuously. This study measures the state, not the decay rate; the checked subset's skew toward actively used records suggests the unchecked 38.8% would fare somewhat worse, and that is stated as expectation, not measurement.
In practice
An SDR planning a cold-email sequence from raw list counts should divide by three and a half, not by two: 100,000 contacts at this database's observed rates yields roughly 53,000 addresses and roughly 28,000 verified-deliverable ones (derived). Bounce budgets and domain-reputation planning should assume the 14.60% hard-bounce rate applies to any unverified list segment.
A founder evaluating "95% deliverability guaranteed" claims from list vendors now has a measured base rate to price them against: a maintained database's own checks came back 53.98% verified. A guarantee that outruns that by forty points is describing a heavily pre-filtered subset, and the honest question is what fraction of the raw list survived the filter.
Limitations
The checked subset is selected. Addresses were verified as they were used, so the 655,442 skew toward records campaigns touched. The unchecked 415,510 addresses are probably worse, not better; no figure here should be extrapolated to them without saying so.
Verification is point-in-time. A verified address decays after its check. The statuses reflect each address's most recent check, not a single synchronized sweep.
Catch-all is a ceiling and a floor. Some catch-all addresses deliver fine; some silently discard. The 18.05% cannot be split further by any check we ran.
One database, one verification pipeline. Different sourcing and different verifiers will move these shares by points. The claim is the measured stack, not universality.
Email presence itself is covered elsewhere. The 52.57% presence figure and its context live in the database audit; this study starts where that one ends.