Article
Why Do Email Verifiers Give Different Results for the Same Address?
By Unlimited Verifier Team ·

Summary
Different email verifiers give different results because each runs its own SMTP mailbox probe, with its own sender identity, timeouts and retry rules, then maps the server's reply onto its own labels. Most disagreement appears at uncertain servers such as catch-all, greylisted and slow domains. Testing each tool against a labelled sample of your own list reveals more than any published accuracy claim.Why do email verifiers give different results for the same address?
Different verifiers return different results for one address because each runs its own SMTP mailbox probe, with its own sender identity, timeouts and retry rules, then maps the server's reply onto its own labels. The disagreement concentrates in a narrow band of uncertain servers. For the wider picture, see our email verifier accuracy guide.
Quick answer
- Probing differs. Verifiers use different sender identities, timeouts and retry rules, so the same server can answer each one differently.
- Uncertain servers drive most conflicts. Catch-all, greylisted and slow servers can't be settled by a single probe, and that's where results diverge.
- Labels differ. One tool's "Risky" may be another's "Unknown" or "Valid", so compare category definitions before headline rates.
- Most conflicts are small. In our illustrative 10,000-row model, 4.7% of rows changed category, and only 0.4% flipped between Valid and Invalid.
- Test on labelled data. A 500-row sample with known outcomes tells you more about your list than any published accuracy figure.
The useful question isn't which tool is "right" in the abstract. Each one answers a slightly different question, and the gap shows up at the edges, where the server can't give a clear yes or no.
Bottom line: Different results usually mean different questions and different labels, not one tool being simply right and the other wrong.
What happens during an SMTP mailbox check?
A mailbox check is a short exchange with the recipient's mail server, and each step leaves room for tools to differ. The exchange follows the Simple Mail Transfer Protocol defined in RFC 5321.
- Syntax check. The address is tested against format rules. Malformed strings never reach the network.
- Domain and MX lookup. The verifier confirms the domain exists and finds its mail exchanger (MX) records.
- Connection and greeting. It opens a session on port 25 and identifies itself with HELO or EHLO.
- Sender and recipient commands. It sends
MAIL FROMwith a sender address, thenRCPT TOwith the address under test. - Reading the reply. A
250means the server will accept mail for that recipient. A550with enhanced code5.1.1means the mailbox doesn't exist, as defined in RFC 3463. A4xxis a temporary deferral. - Disconnect before DATA. A well-behaved probe quits before any message body is sent, so nothing is delivered.
Step 5 is where the trouble starts. The reply is advisory, not a guarantee. Some servers accept every recipient at this point and filter later. Others defer unfamiliar senders, and some refuse probes outright. Each verifier has to decide what those replies mean, and those decisions cause most differences.
Bottom line: A verifier sees one server reply to one command, so its label is an interpretation, not a fact.
Which factors make verifiers disagree?
The eight factors below are the usual causes of conflicts. Each one changes a result in a specific way, and each can be checked on your own list.
| Factor | What it does to the result | Why tools differ | How to check |
|---|---|---|---|
| Probe identity (sender address, HELO name, IP reputation) | Some servers refuse or defer unfamiliar probers | Each vendor runs its own sending infrastructure | Re-run a sample through a second tool |
| Timeouts | A slow but valid server returns "Unknown" | Timeout length and retry count are set per vendor | Count Unknowns on domains with slow MX records |
| Greylisting | First contact gets a temporary 4xx |
Retry policy decides whether a deferral becomes a result | Re-verify deferred addresses after a delay |
| Catch-all detection | Accept-all servers look Valid to a naive probe | Some tools test a random address on the same domain, others don't | Check whether Valid results sit on catch-all domains |
| Rate limiting | Large bursts from one IP get throttled | Vendors use different IP pools and pacing | Compare a small batch against a large one |
| Category mapping | Same server reply, different label | Each vendor defines Risky, Unknown and Valid differently | Read each tool's category definitions |
| Auxiliary flags (disposable, role, spam-trap) | Flags change the label without changing validity | Disposable and blocklist data vary and age | Compare flags on known disposable and role addresses |
| Time and caching | Yesterday's result may not match today's | Platforms keep results for different windows | Export results on the day you run them |
Two factors matter most. Category mapping comes first. Verifiers use different status names and sub-codes, so one tool's "risky" isn't guaranteed to equal another's. Check each vendor's documentation for what it counts as risky, unknown or disposable. Our guide on why ZeroBounce and NeverBounce disagree on the same email shows how this plays out for one pair.
Catch-all handling comes second. A verifier that tests a random nonexistent address on the same domain can learn whether the server accepts everything. A tool that skips that test may label every address on that domain Valid. Neither answer is a lie. They're different confidence levels presented under the same word.
Bottom line: Before comparing two tools' numbers, check how each defines its labels and whether it tests for catch-all servers.
How much disagreement is normal?
There's no published benchmark that sets a normal level, so we use a working rule: treat disagreement above about 5% of rows as a reason to investigate. The model below shows how a 4.7% disagreement breaks down. It's illustrative, not a measured result.
Worked example: two tools, one 10,000-row list
Tool A and Tool B both verify the same 10,000 addresses:
| Tool A \ Tool B | Valid | Invalid | Risky | Row total |
|---|---|---|---|---|
| Valid | 8,020 | 0 | 310 | 8,330 |
| Invalid | 40 | 860 | 0 | 900 |
| Risky | 0 | 120 | 650 | 770 |
| Column total | 8,060 | 980 | 960 | 10,000 |
The arithmetic:
- Exact agreement: 8,020 + 860 + 650 = 9,530 rows, or 95.3%.
- Total disagreement: 10,000 − 9,530 = 470 rows (4.7%).
- Headline rates: Tool A reports 83.3% Valid (8,330 ÷ 10,000) and Tool B reports 80.6% (8,060 ÷ 10,000). The Valid gap is 2.7 points, which hides 470 rows that moved.
- Hard conflicts (Valid vs Invalid): 40 rows, or 0.4%. Tool A rejects them and Tool B accepts them.
- Soft conflicts (Valid vs Risky): 310 rows, or 3.1%. These are most likely catch-all addresses, where one tool is confident and the other isn't.
- Risky vs Invalid: 120 rows, or 1.2%.
The practitioner's point: the 40 hard conflicts matter most, because sending to them could mean mailing addresses one tool considers dead. The 310 soft conflicts are mainly a labelling question, which you can settle with engagement data. Review in that order: hard conflicts, then Risky vs Invalid, then soft conflicts.
Bottom line: A small headline gap can hide hundreds of row-level changes, so compare the cross-tab, not the summary percentages.
Where do verifiers disagree most often?
Verifiers disagree most where the server can't give a clear yes or no. Three patterns recur.
- Catch-all business domains. Many corporate gateways accept recipients during the SMTP session and filter afterward. The label depends on whether the tool detects the catch-all setup. See our catch-all accuracy guide.
- Large consumer providers. These providers tend to rate-limit probing and don't always reveal mailbox existence in a way every tool reads the same. The Gmail vs Outlook accuracy comparison covers these patterns.
- Slow or greylisted small hosts. Budget and self-hosted servers often defer first contact or respond slowly. Whether that becomes "Unknown" or a final answer depends on the tool's retry policy.
Your domain mix matters more than any tool's reputation. Teams running email verification for ecommerce and SaaS workflows often see more Risky results on B2B lists, because corporate domains concentrate catch-all setups. That's expected behaviour, not a sign that one tool is broken.
Bottom line: Disagreements cluster on catch-all, slow and rate-limited domains, so check your domain mix before blaming a tool.
What should you check before comparing verifiers?
Five questions reveal most of the differences above. The third column shows what Unlimited Verifier documents, so you can compare it with any other tool on the same terms.
| Question to ask | Why it changes results | Unlimited Verifier's documented position |
|---|---|---|
| How are results categorised? | The same reply can get a different label | Sorted into Valid, Risky or Invalid; catch-all, disposable and role-based addresses are flagged |
| Is catch-all detection offered? | Accept-all servers look Valid to naive probes | Detects catch-all domains; see the catch-all email verification page |
| How long are results kept? | Old exports may not match current server behaviour | Free: 24 hours. Pro: 30 days |
| Can you inspect the logic? | Lets you see how probes and labels are built | Verification engine published as open source on GitHub |
| How are bulk runs handled? | Large batches can be throttled or split | API batches of up to 10,000 emails per request, on Pro (100,000 API emails a month) |
Bottom line: A verifier you can audit, and whose categories you can read, is easier to compare than one that offers only a headline rate.
How should you reconcile conflicting results?
Decide from the pattern of results across tools, not from any single label. The table covers the common cases.
| Result pattern | Most likely meaning | Recommended action |
|---|---|---|
| Invalid from two or more tools | Mailbox or domain doesn't exist | Suppress permanently |
| Valid from two or more tools, recent engagement | Likely deliverable | Send in your main segment and monitor |
| Valid in one tool, Invalid in another | Disputed; one probe was likely deferred or refused | Re-verify after a delay, then check engagement |
| Catch-all flag in one tool, Valid in another | Accept-all server; mailbox can't be confirmed | Hold for late-stage sends; prioritise by engagement |
| Unknown or timeout only | Server didn't answer clearly | Retry later; don't treat as Invalid by default |
| Disposable or role flag | A policy question, not a validity question | Decide by your compliance and sending policy |
Apply the table in four steps:
- Resolve hard conflicts first. Re-verify Valid-vs-Invalid addresses after a delay, then suppress any that stay Invalid.
- Segment soft conflicts by engagement. Send to recent engagers first, and hold catch-all and disposable addresses for later.
- Check your Unknown share. A high share of Unknowns points to a timeout or greylisting problem, not bad data. Our unknown rate guide explains what to expect.
- Write down your rules. Record the thresholds and category decisions, in case an ESP or auditor asks.
Bottom line: Agreement across independent tools is strong evidence, a single disputed result is weak evidence, and engagement data breaks ties.
How do you test verifiers on your own list?
A labelled sample is the only method that shows how a tool performs on your addresses rather than on a vendor's test set.
- Build a labelled sample of about 500 addresses. Use 200 known-good addresses (recent openers or confirmed customers), 200 known-bad addresses (hard bounces from the last 90 days, which are more likely to still be current), and 100 ambiguous addresses (catch-all, role and disposable).
- Store the true outcomes separately. Keep a labels file keyed by email address, and upload only the addresses.
- Plan the uploads. On the Free plan each list holds up to 150 emails, so 500 addresses need four lists (150 + 150 + 150 + 50). That uses four of the 25 lists available each month.
- Upload each file. Use the upload and verify a list page. CSV and Excel files up to 10MB are accepted, and the email column is detected automatically. The Free plan has no API access, so this test runs by upload.
- Run deep verification on the ambiguous set. Free includes 750 deep verifications a month, so the 100 ambiguous rows use 100 of them.
- Export results the same day. Free results are kept for 24 hours, so save them as CSV, Excel or JSON before they expire.
- Join results to your labels and count. Count known-bad addresses marked Valid (false-valid, the most dangerous error for a sender) and known-good addresses marked Invalid (false-invalid).
- Repeat with a second tool and compare. Run the same file through another verifier, then compare both rates side by side.
Illustrative arithmetic: if Tool X marks 4 of 200 known-bad addresses as Valid, its false-valid rate is 4 ÷ 200 = 2.0%. If Tool Y marks 11 of them as Valid, its rate is 11 ÷ 200 = 5.5%. On this sample, Tool X is the safer choice for sending, even if Tool Y shows a higher Valid rate on your general list. For the terminology, see our false positives explainer, since the term is used in both directions.
Once you've picked a tool, schedule re-tests using the list maintenance best practices cadence.
Bottom line: A 500-row labelled test gives you false-valid and false-invalid rates for your own data, which beats any headline accuracy figure.
Are vendor accuracy claims a reliable guide?
Treat vendor accuracy figures as claims measured against each vendor's own ground truth, not as a common yardstick. The sample, the labels and the definition of "accurate" all vary, so the same percentage can mean different things. Our vendor accuracy claims article explains how to read them, and the independent accuracy benchmark covers third-party comparisons.
Unlimited Verifier states 99.5% accuracy for its deep verification. That's the company's own figure, not an independent benchmark, so use it as a starting point and then run your own labelled test. For the category-wide 99% claim, see do email verifiers really have 99% accuracy.
Price is a weak proxy for accuracy too. The email verification pricing entry explains how to compare models. For Unlimited Verifier, the arithmetic is simple. Pro includes 25,000 deep verifications a month, so the cost per 1,000 deep checks at full use is:
- Monthly billing: $29 ÷ 25 (25,000 ÷ 1,000 = 25 blocks) = $1.16 per 1,000
- Annual billing: $25 ÷ 25 = $1.00 per 1,000
Standard verification is unlimited on every plan with no per-email fees, so those figures cover only deep checks.
Bottom line: Judge a verifier on a labelled test of your own list, not on its published accuracy figure or its price.
What are the most common mistakes when comparing verifiers?
- Comparing headline percentages instead of row-level results. Two tools can report similar Valid rates while disagreeing on hundreds of individual rows.
- Treating Risky as Invalid. Catch-all and disposable addresses belong in their own bucket, not the dead-address list.
- Testing on a list you've already cleaned. Once bad addresses are removed, the remaining sample can't show how tools handle bad data.
- Trusting one large batch. Rate limiting may degrade results in big bursts, so spot-check a smaller sample.
- Forgetting that results expire. Free results last 24 hours and Pro results 30 days, so re-verify before relying on older exports.
- Ignoring greylisting. A deferred result on the first pass isn't final. Retry before classifying the address.
- Assuming Valid means delivered. A Valid label says the server didn't reject the address at that moment. Inbox placement also depends on engagement and complaints, as set out in Google's email sender guidelines. See does a valid result mean the email will be delivered.
- Using verification as proof of consent. Verification can't show whether someone opted in. That's a compliance and hygiene question.
Bottom line: Most comparison errors come from reading summary numbers too literally, not from the tools themselves.
How we compared (and limitations)
This article is based on public documentation: the SMTP and enhanced status code RFCs, Google's sender guidelines, and Unlimited Verifier's published plan facts and open-source engine. We haven't run a controlled benchmark across verifiers. The 10,000-row cross-tab and the 500-row test figures are illustrative models that show the arithmetic, not measured results.
Verifier behaviour changes as mail servers and tools update, so re-test periodically instead of relying on one snapshot. Unlimited Verifier's 99.5% deep-verification accuracy is the company's stated figure, and we haven't independently verified it. Because the engine is open source on GitHub, you can check its logic directly.
Bottom line: Use the models to understand the pattern, then measure on your own list.
Fair comparison steps
- Assemble 500+ addresses with known outcomes from past bounces and confirmed opt-ins.
- Run each verifier on the identical file, same day.
- Map every tool's labels to Valid, Invalid and Uncertain.
- Calculate agreement and false-positive rates per server type (catch-all, greylisting, normal).
- Choose the tool whose errors fall where your risk is lowest.
Frequently asked questions
Is one email verifier more accurate than another?
Not in the abstract. Accuracy depends on the list, the mix of server types and the label definitions a tool uses. The most reliable way to judge is to test several verifiers on a labelled sample of your own addresses and measure agreement with known outcomes.
Why does a catch-all domain return Risky or Unknown?
A catch-all server accepts mail for any recipient at the SMTP stage, so a probe cannot confirm that a specific mailbox exists. Verifiers label these addresses differently, which is one of the most common reasons results conflict.
What is greylisting and how does it affect verification?
Greylisting makes a server temporarily reject mail from unfamiliar senders with a 4xx reply, asking them to retry later. Verifiers that do not retry may label the address uncertain, while others retry and reach a firmer result.
Does an SMTP 250 response guarantee the mailbox receives mail?
No. A 250 reply at RCPT TO means the server is willing to accept mail for that recipient at that moment. Servers can still filter, bounce or drop messages later, so the reply is advisory rather than a delivery guarantee.
What does a 550 5.1.1 reply mean?
A 550 reply with enhanced status code 5.1.1 indicates that the recipient mailbox does not exist, as defined in RFC 3463. Most verifiers treat this as invalid, though some servers return it inconsistently.
Why do verifiers use different sender addresses?
Each vendor sets its own MAIL FROM address, HELO name and IP reputation. Some servers treat unfamiliar or poorly reputed senders differently, so the same mailbox can produce different replies depending on who is asking.
Are the differences between verifiers large?
Usually they are small. In an illustrative 10,000-row model, about 4.7% of rows changed category between tools, and only about 0.4% flipped between Valid and Invalid. Your own list may differ, so measure it directly.
How do I compare verifiers fairly?
Build a sample of 500 or more addresses with known outcomes, run each tool on the same file, map every label to a common scale, and compare agreement rates and false positives separately for each server type.
Should I remove addresses marked Risky or Unknown?
It depends on your sending goals and risk tolerance. Many senders suppress Invalid addresses immediately and handle Risky or Unknown ones more cautiously, such as sending a low-volume re-engagement test before adding them to full campaigns.
Cut Bounces and Protect Your Sender Reputation With Unlimited Email Verification
Every invalid address you send to hurts your sender reputation and your inbox placement. Verifying your list before each send means fewer bounces, healthier deliverability, and a predictable cost. With Unlimited Verifier, standard verification is unlimited on every plan, with no per-email fees.
Unlimited Verifier checks syntax, domain, MX records and the SMTP mailbox. It also detects catch-all domains and flags disposable and role-based addresses. Export a CSV from your email platform or CRM, upload and verify the list, then re-import the clean results. You can also automate checks with the email verification API.
| Free | Pro | |
|---|---|---|
| Price | $0 | $29/month ($25/month billed annually) |
| Lists | 25 a month, up to 150 emails each | Unlimited, up to 25,000 emails each |
| Deep verifications | 750 a month | 25,000 a month |
| API access | No | Yes (100,000 emails a month) |
| Result retention | 24 hours | 30 days |
Why teams choose Unlimited Verifier
- Unlimited standard verification on all plans, including Free, with no per-email fees.
- Catch-all detection so you know which domains accept every address.
- Clear Valid / Risky / Invalid results, exportable as CSV, Excel or JSON.
- An open-source verification engine you can inspect on GitHub.
Unlimited Verifier states 99.5% accuracy for deep verification.
Create a free account and verify your next list before you hit send.
Want to test a single address first? Try the free email checker.