What should I look for in an email verification service?

Ask how catch-all domains are reported, whether you are billed for results nobody could determine, whether the price is shown before the spend, whether anything is sent to the addresses, where the data is processed and how long it is kept, and whether credits expire. Those six answers separate services far more than headline accuracy figures do.

Why do two email verification services give different results for the same list?

Mostly because they treat undetermined addresses differently. Catch-all domains accept every address, so mailbox existence cannot be established from outside. One service reports those as valid, another as invalid, a third as undetermined — same list, same facts, three different answers.

Is a 99% accuracy claim meaningful?

Only if you know what it counted. Accuracy figures are easy to inflate by guessing at the addresses nobody can settle and scoring the guesses as correct. Ask what happened to the catch-all domains in the sample: they were either excluded from the calculation or guessed at.

Should I verify a list with a service or clean it manually?

Manual cleaning catches typos, duplicates and malformed rows, and nothing else. It cannot tell you whether a correctly spelled address still has a mailbox behind it, which is what causes hard bounces. Do the manual pass for free on any size of list, and use a service for the question a spreadsheet cannot answer.

How to choose an email verification service

Eight questions that separate them, and why the accuracy figure on the homepage is not one.

Every service in this market advertises a number in the high nineties, and those numbers are not comparable — they do not count the same things. What follows is the set of questions that actually distinguish one from another, written so you can put them to anybody, including us.

You will notice we do not name competitors or quote their pricing. Comparison pages that do are almost always advertisements, and the figures in them were true once and are repeated forever. We cannot verify anyone else's current terms, so we do not assert them.

The eight questions

1. How are catch-all domains reported?

This is the question that separates services most, and the one least often asked. A catch-all domain accepts mail to every address, real or not, so mailbox existence cannot be established from outside — by anyone. What differs is what a service does about it: report it as valid, report it as invalid, or say so.

Ask to see a result for a catch-all domain before you buy.

2. Am I billed for results nobody could determine?

Catch-all and could-not-complete outcomes are the two where the work was done, the mail server was asked, and it declined to say. Charging for those is charging you for the service’s own inability to find out, and it quietly rewards guessing.

Ask whether undetermined results are refunded or billed.

3. Do I see the price before anything is spent?

A list is not a fixed cost until duplicates and malformed rows have been counted. Quoting after the run means the number that mattered arrived too late to decline.

Ask whether a quote is shown, and whether it is binding.

4. Is anything sent to the addresses?

A verification check should open a conversation with the mail server and stop before a message exists. Any service that actually delivers mail to confirm an address is mailing your customers without their knowledge, using your list.

Ask whether a message is ever delivered. The answer should be an unqualified no.

5. Where is my list processed, and how long is it kept?

You remain the controller of those addresses. If the service is a processor, you need to know the jurisdiction, the retention period, whether you can shorten it, and whether a data processing agreement exists at all.

Ask for the retention default in days, and for the DPA.

6. Do credits expire?

Most of this market prices as a monthly allowance that resets whether you used it or not, which suits a service more than a shop that cleans a list twice a year.

Ask what happens to unused credits at the end of a month.

7. What does the accuracy figure actually count?

A high number is easy to produce if undetermined results are quietly reported as valid — every guess becomes a billable correct answer, and the cost lands later on your bounce rate. The figure means nothing without knowing what it excluded.

Ask what happened to the catch-all domains in the sample.

8. Can I get the results out in a usable shape?

A verdict per address is only useful if you can act on it: the clean list to import, the original annotated so you can check the reasoning, and the removed rows so nothing disappears without a trace.

Ask what the download actually contains.

Why the same list gets different answers

Run one list through two services and the counts will disagree, sometimes badly. That is usually not one of them being wrong. It is catch-all domains being handled differently.

On a catch-all domain the mail server accepts every address put to it, real or invented, so there is no observable difference between a live mailbox and an imaginary one. Every service faces the same wall. One calls those addresses valid, one calls them invalid, one says it could not tell — same list, same facts, three different answers, and only the third is describing reality.

This is also why the highest reported "valid" count is the worst signal to buy on. The service that guesses generously looks best on the summary screen and worst on your bounce rate three weeks later.

Do you need a service at all?

A fair question, and the honest answer is that some of the work is free.

What a manual pass catches: typos you can see, duplicates, malformed rows, obviously fake entries, and addresses at domains that plainly do not exist. A spreadsheet, an hour, and some patience will find all of it at no cost.

What it cannot catch: whether [email protected] — spelled correctly, at a real company, looking perfectly ordinary — still has a mailbox behind it. Nothing about that row is visibly wrong. It is also exactly the row that hard bounces and damages your sending reputation, because people leave jobs and abandon addresses without editing your list.

So: do the manual pass, on any size of list, and use a service for the question a spreadsheet cannot answer. If your list is a few hundred addresses you collected yourself last month, the manual pass may genuinely be enough.

Where this leaves you

Take the eight questions to whoever you are considering. If a service answers them plainly, you can judge it. If it answers with an accuracy percentage, you have learned something too.

For what it is worth, here are our own answers: catch-alls come back marked as undetermined and are not charged for; neither are results where the check could not complete; the price is shown before anything is spent and duplicates and malformed rows are never billed; nothing is ever delivered to an address being checked; lists are processed in the European Union and kept 90 days by default, which you can shorten or erase; and credits do not expire. ListCheckup opens in the coming weeks. Leave your address and we will tell you the day it does.