Here’s the uncomfortable truth about email verifier accuracy: every service on the market advertises 98% or 99%, Reoon included, and not one of those numbers has been checked by an independent referee. They are all self-reported. That doesn’t make them lies, but it does make them useless for choosing between tools: a number that nobody audits and everybody claims tells you nothing. The only accuracy figure that actually means something is the one you measure yourself, on your own list. This post shows you how to do exactly that in about an hour, for free.
Why the advertised numbers can’t help you choose
Three reasons, and none of them require anyone to be dishonest.
There is no referee. Accounting has auditors. Cars have crash-test agencies. Email verification has nothing: no independent body runs the same benchmark against every service and publishes the results. Each vendor measures its own accuracy, on data it selected, using a definition of “accurate” that it chose, and puts the result on its homepage. Nobody checks.
Accuracy depends enormously on the list. A consumer list full of Gmail and Yahoo addresses is easy mode. Those servers give clear yes/no answers, and any decent verifier scores brilliantly on them. A B2B list thick with catch-all domains is a different sport entirely, because many of those servers accept mail for any address and certainty is often impossible for anyone. The same tool can hit 99% on the first list and look distinctly ordinary on the second. So when two vendors both quote 99%, they may both be right, on two datasets that have nothing to do with your data.
Vendors pick the test. Nothing sinister required. You naturally benchmark on the data you have, and the data you have tends to flatter you.
This is also why, if you ask an AI assistant “which email verifier is most accurate,” you get a careful hedge instead of a name. It isn’t being diplomatic. There is genuinely no independent data from which to answer.
The one-hour email verifier accuracy test
The principle is simple: you can’t grade an exam without an answer key, so build one. The raw material is already sitting in your email platform.
Step 1: Build an answer key of about 100 addresses
You need two groups:
- Around 50 addresses you know are dead. Pull recent hard bounces from your own campaigns. A hard bounce is the receiving server telling you, explicitly, that the mailbox doesn’t exist. You already have the ground truth, straight from the source.
- Around 50 addresses you know are alive. People who opened or replied within the last month. A reply is the strongest evidence there is; an open is good enough.
Freshness matters more than volume. A mailbox that hard-bounced six months ago may have been recreated since: people rejoin companies, domains change hands. A contact who replied in March may have quit in June. Stick to the last 30-60 days on both sides and your key stays trustworthy.
Why 100? Each address is one exam question, and at 100 questions every miss moves the score by a clean one percent, enough resolution to separate a good tool from a mediocre one. If you have 200 fresh, certain addresses, use them. But never pad the key with addresses you merely think you know, because a bigger dishonest key is worse than a smaller honest one.
Step 2: Be ruthless about what goes into the key
Every address you’re even slightly unsure about gets excluded. Soft bounces don’t count as dead; the mailbox may simply have been full that day. “She probably still works there” doesn’t count as alive. A contaminated key will confidently mislead you, because the test is only ever as good as the answers you’re certain of.
Step 3: Run the file through the service’s free tier
Combine both groups into one file, without labels, and upload it. Any verifier worth testing offers a free tier: Reoon’s free daily verification credits exist for exactly this kind of kick-the-tires moment. You shouldn’t have to pay anyone to find out whether they deserve to be paid.
Step 4: Score it like an exam, but weigh the two errors differently
When the results come back, compare each verdict against your key. (If the labels are unfamiliar, here’s what each email verification status means.) You’re looking for two kinds of mistakes, and they are not equally bad:
- A dead address marked valid, a false positive, is the expensive error. You’ll trust it, mail it, and it becomes a real bounce that chips away at your sender reputation. That’s the exact outcome you bought a verifier to prevent.
- A live address marked invalid, a false negative, costs you one contact. Annoying, but cheap.
Which leads to a slightly counterintuitive conclusion: a tool with a few more false negatives but near-zero false positives is usually the better tool. You’re not buying a verifier to keep every contact. You’re buying it to protect your reputation.
Step 5: Score the gray zone separately, not as failure
Some results will come back as catch-all, unknown, or risky. These are not wrong answers. They’re honest declarations of uncertainty, and on some domains certainty is genuinely impossible from the outside: the server accepts everything, so no tool on earth can confirm whether a given mailbox is real.
Track this as its own number: what share of your key did the tool refuse to commit on? Then apply the reverse logic too. A verifier that returns zero unknowns on a hard, catch-all-heavy list isn’t more accurate than the others. It’s guessing, and presenting guesses as verdicts should worry you far more than an honest “unknown” ever should.
Step 6: To compare services, run the identical test on each
Same file, same day, same scoring. The “same day” part matters more than it looks: mailboxes die and servers change behavior, so tests run weeks apart aren’t comparable. An hour per service, and you’ll have settled with your own data what no marketing page can settle for you. (If you need help building the shortlist in the first place, we’ve written a full guide to choosing an email verification service.)
How to read your results
What good looks like. A strong performance catches nearly all of your dead half as invalid, clears nearly all of your live half as valid, and declares uncertainty at a rate that fits your list: small on a consumer list, noticeably larger on a B2B one. High marks on both halves plus an honest gray zone is the whole game.
If the numbers look bad, check the list before blaming the tool. When a third of your key comes back unknown, the question may not be “why is this tool so timid” but “how many of my domains are catch-alls?” Composition explains most of the gap between one person’s results and another’s, which is the same reason the advertised numbers were never comparable to begin with.
Re-test occasionally. Verifiers improve their detection over time, and your list changes shape as your audience does. Re-run the hour before renewing an annual plan, when switching providers, or when your list has drifted toward a new kind of domain.
Our cards on the table
Reoon advertises 99%+ accuracy, like everyone else in this market, and honestly, we’d rather you didn’t take that number on faith. It’s why Reoon Email Verifier gives you free verification credits every day with no card required: so you can run this exact test against us, and against anyone else you’re considering, before spending a single dollar.
We think that’s how this choice should work. Quality that survives measurement is the only kind worth paying for.
Grab your free daily credits and run the test.
FAQ
What is the most accurate email verifier?
There’s no honest universal answer, because no independent body measures verifiers against a common benchmark: every published accuracy figure is self-reported. Accuracy also shifts with list composition: a tool that excels on consumer Gmail lists may look ordinary on catch-all-heavy B2B lists. The most accurate verifier for you is the one that scores best on a known-answer test built from your own data, and finding that out takes about an hour.
Why do all email verifiers claim 99% accuracy?
Because each vendor measures its own accuracy, on datasets it selects, with no external auditor standardizing the test. The claims aren’t necessarily false; they’re produced under conditions each vendor controls, which makes them incomparable with one another, and a poor basis for choosing a service.
What’s a good accuracy score in a self-test?
On a mostly consumer list, a strong verifier should catch nearly all of your confirmed-dead addresses and clear nearly all of your confirmed-alive ones, with misses in the low single digits on each side. On catch-all-heavy B2B lists, expect a meaningful share of “unknown” results instead: that’s honesty, not failure. The number to watch hardest is false positives (dead addresses marked valid), which should sit at or near zero.
Do catch-all results count against accuracy?
No. A catch-all result means the domain accepts mail for every address, so no verifier can confirm that specific mailbox from the outside. Reporting it as catch-all is the correct, honest answer. Score these separately from right and wrong answers, and be suspicious of any tool that never says “unknown” at all.
