Vendor accuracy figures and independent tests disagree for a structural reason. A vendor's headline number is usually measured on the rows that got a clear valid or invalid, after the hard rows are set aside. An independent test like Hunter's scores every address, including catch-all and unanswered ones. Both numbers can be honest, and only the second tells you how your whole list will fare.
NeverBounce's own API shows where the hard rows go. Its single-check endpoint returns five codes: valid, invalid, disposable, catchall and unknown (developers.neverbounce.com, read 24 September 2026). The docs describe catchall as accept-all, or unverifiable, and unknown as a server that can't be reached. Our page on what accept-all unverifiable means goes deeper on that code.
So the useful question about NeverBounce, or any verifier, has two parts. What share of your list got a definite verdict? And how often was that verdict right? The first is coverage and the second is accuracy. A B2B list shows why coverage matters: about 28% of a 10,000-email B2B benchmark sat on catch-all domains, where no SMTP check can confirm a mailbox.
You can test this yourself in an afternoon. Take 500 addresses whose fate you know: recent sends that delivered, and addresses that hard bounced. Run them through the tool, then count the rows it answered and the rows it got right. Run the same file through a second verifier and you'll see why verifiers disagree on the hard rows.
ZapBounce reports both numbers for every list. Each row comes back valid, invalid, catch_all or unknown, the summary shows how much of the list reached a verdict, and unknowns are never billed. The accuracy page explains the method, and 1,000 checks cost $5.
Related questions
Is NeverBounce accurate?
NeverBounce is accurate on the addresses its SMTP check can get a clear answer for, like any verifier. Catch-all domains and servers that block probes land in its catchall and unknown codes. Hunter's November 2025 benchmark of 15 tools, NeverBounce included, measured the field at 31.2% to 70.0% overall.
Why does NeverBounce return so many catchall results?
Catch-all domains accept every address at the SMTP step, so no verifier can confirm a single mailbox there. About 28% of a 10,000-email B2B benchmark sat on catch-all domains. Business lists land in that code more than consumer lists.
How do I check a verifier's accuracy myself?
Run 500 addresses with known outcomes through the tool: delivered sends and hard bounces. Count how many rows got a verdict, then how many verdicts matched reality. Those two counts are coverage and accuracy.
Does NeverBounce charge for unknown results?
NeverBounce's pricing and billing pages blocked automated reads on 24 September 2026, so check its current rule on its own site. ZapBounce never bills an unknown result or a duplicate.