Our data · snapshot of 5 October 2026
How many B2B domains are catch-all? We measured 1.46 million of them
Across 4,687,209 work-email verifications covering 1,460,664 distinct company domains, 31.90% of the addresses and 26.49% of the domains came back catch-all — a server that accepts mail for any address at all, so no verifier on the market can tell you whether that particular mailbox exists. Just over a quarter of B2B domains behave this way. That one fact explains most of what looks like disagreement between email-verification tools.
Every verdict, with the counts behind it
These are raw counts from one pass over our verification table on 5 October 2026. Percentages are of all 4,687,209 rows.
| Verdict | Addresses | Share |
|---|---|---|
| Deliverable — mailbox confirmed | 2,824,842 | 60.27% |
| Catch-all — server accepts anything | 1,495,369 | 31.90% |
| Undeliverable — rejected | 330,192 | 7.04% |
| Risky | 14,051 | 0.30% |
| Code not in our own enum — we do not know what it means, and we are not going to guess | 22,755 | 0.49% |
Counted by domain rather than by address, 386,860 of 1,460,664 domains (26.49%) returned a catch-all verdict at least once. The two figures differ because catch-all domains tend to be larger: the same domain accounts for many addresses.
What this means if you are buying an email tool
A vendor’s accuracy claim and your bounce rate can both be true
On roughly a third of addresses there is nothing to be accurate about. A catch-all server answers yes to every mailbox, real or invented, so a verifier has no signal to work with. Vendors differ mainly in what they then say: some return "valid", some "risky", some "accept-all", some quietly drop the row. Same reality, four different labels — and the one that says "valid" looks the most accurate right up to the moment you send.
Compare tools on the same list, not on each other's numbers
Every published accuracy figure, ours included, is measured on the vendor's own traffic. Take 50 rows from the segment you actually sell to, run them through each tool, and count two things: how many came back with an address at all, and how many of those were confirmed rather than labelled catch-all. The second number is the one that predicts your bounce rate.
Catch-all is not the same as bad
A catch-all domain is usually a normal company that configured its mail server to accept everything. Plenty of those addresses deliver. What you lose is the ability to know in advance — so these rows belong in a smaller, more careful send, not in the bin and not in a blast.
How the number was produced, and where it stops being reliable
Published in full because a statistic without its limits is just a marketing claim.
- Snapshot date: 5 October 2026. One pass over our verification table, no sampling, no extrapolation.
- What the population is: addresses that passed through our own pipeline because a customer asked for them. This is not a random sample of the web, and B2B prospecting skews it towards companies people sell to.
- Dating is incomplete: 75.90% of the rows carry no verification timestamp. The dated subset is 1,129,899 checks between 19 March and 5 October 2026.
- Single-probe verdicts: for 928,576 domains (63.57%) the verdict rests on one check. A domain can change its configuration.
- What we deliberately do not publish: the share of role-based, free-provider and disposable addresses, and the mail-provider breakdown. Those columns are empty or barely populated in our data, so any figure we printed would be invented rather than measured.
Reusing these figures is fine — a link back to this page is appreciated so readers can see the date and the limits above.
Why we have this data, and what our own numbers look like
Scrupp reads the LinkedIn or Sales Navigator search you already have open and returns those rows with a work email and phone attached, verifying each address at export time — which is why we hold millions of verification verdicts rather than a stored list someone else built. Measured across all exports, an address comes back for 36.3% of rows and 10.0% are confirmed by an accepted delivery; rows where nothing is found are not charged. The gap between those two figures is largely the catch-all share on this page: we will not claim a delivery we cannot observe.
Looking for the definition rather than the numbers? See what a catch-all email is, or the practical side of finding work emails and finding emails from LinkedIn profiles.