The email verification metric that almost nobody measures

Every email verification tool reports what it caught. Invalid addresses removed, bounces prevented, the percentage of a list cleaned. Those numbers land in a dashboard, they look like performance, and they are the basis on which most businesses decide whether the tool is doing its job.

No verification tool reports what it missed. A real, reachable contact that a tool failed to confirm does not produce an error, an alert, or a line item anywhere. It is simply absent from the output file, and nothing about that file suggests it should have been there.

The result is that businesses evaluate verification on one half of its performance, the half that happens to be easy to measure. The invisible half is frequently the more expensive one.

Two kinds of mistake, only one of which is visible

A verification tool can get things wrong in two directions.

It can say an address is good when it is not. That produces a bounce, which generates a notification, appears in deliverability reporting, and feeds into how mailbox providers score the sending domain. It is a loud failure. Teams notice it, vendors know teams notice it, and accuracy claims across the industry are built around minimising it.

It can also fail to confirm an address that was genuinely reachable. That produces nothing. The contact is dropped from the usable list or parked in an ambiguous category, the campaign goes out to a smaller audience than it should have, and the outcome looks like a clean list rather than an incomplete one. There is no error state for a person who was never found.

The asymmetry in what these mistakes cost is worth stating plainly. A false positive costs a bounce and a small, recoverable hit to sender reputation. A missed contact is permanent, because nobody goes looking for a record that is not there.

Why this hits business lists hardest

The reason contacts go missing is that a large share of corporate email domains do not answer verification queries plainly.

Some domains are configured to accept mail for any address, so a standard verification check receives a positive response whether or not the mailbox actually exists. Many large organisations also route inbound mail through security gateways that are deliberately configured not to reveal which mailboxes are real, because that information is precisely what an attacker would want.

Checks against these domains come back inconclusive, and inconclusive records get handled inconsistently across the industry. Some tools label them and hand the decision back to the customer. Others resolve them to invalid and move on. Either way, real people end up outside the usable output.

This is not an edge case. Research by Allegrow verifying the primary corporate domains of the largest companies in the United States found that a substantial majority are configured as catch-all, sit behind a security gateway, or both. Only a minority answer a verification query in a way that produces a clear result. Put another way, the problem concentrates exactly where the highest-value contacts sit.

The gap between tools is measurable, and rarely measured

None of this has to be taken on faith. It is testable. It just requires building a dataset where the correct answer is known before any tool touches it, which is considerably more work than most published comparisons are willing to do.

A test of that kind needs two things: real contacts whose reachability has already been established, and invented addresses that could not possibly correspond to a real mailbox, sitting on the same domains and run through every tool on the same day. Without known answers on both sides, a comparison is just measuring how confidently different vendors describe their own uncertainty.

When those tests are run, they tend to reveal something the headline figures do not. Tools that look similar on advertised accuracy can differ substantially on how many real people they actually find, because advertised accuracy is generally built from the visible half of the performance.

Allegrow ran a controlled test of this kind across thousands of addresses on hundreds of enterprise catch-all domains, comparing several verification tools along with the different modes available within them. Two findings are relevant here. Tools separated far more from one another on real contacts found than on false positives, which is the opposite of where the marketing attention goes. And some paid features that promised to resolve ambiguous records did so largely by labelling them invalid, producing tidier output without producing additional reachable people.

That second point is the one worth sitting with, because it is also the question underneath most searches for zerobounce alternatives and their equivalents for every other tool in the category, even when the person searching would not phrase it that way. What they usually want to know is not which tool removes the most addresses. It is which tool leaves them with more of the people they were trying to reach.

What a business is actually losing

The scale of this depends entirely on list size, which is why it goes unnoticed for so long. A small percentage difference in real contacts found is genuinely unimportant on a list of a few hundred. On a list of tens or hundreds of thousands, it is a different conversation.

There are three costs, and none of them appear in reporting.

Reachable prospects nobody ever contacts. They exist, they read email at work, and no salesperson will ever see them, because a verification pass concluded they were not there and no process exists to question that conclusion.

Acquisition spend that never converts to contact. A business that has paid to source a contact record and then loses it to a verification pass has paid for something it will never use. The cost was incurred; the asset was quietly discarded.

A false reading of list quality. A clean list and a complete list are different things, and standard reporting cannot tell them apart. Both look like a low bounce rate.

How to check without taking anyone’s word for it

Any business can run a version of this test on its own data, and its own data is the only sample whose results genuinely apply to it. Published benchmarks describe someone else’s list composition.

Three principles cover most of it. Include contacts whose reachability you already know, such as people who have replied to you, because established contact is the strongest ground truth available. Include addresses that could not possibly be real, so that errors are unambiguous rather than debatable. Run everything on the same list on the same day, with no adjustments between passes.

Then measure the thing that is normally left out. Count how many of the known-real contacts each tool actually found, not just how many bad addresses it removed. That number determines what the business can reach, and it is the one no dashboard will surface on its own.

A final thought

Email verification is sold as a subtraction exercise. It removes bad addresses from a list, so tools get judged on how much they remove and how confidently they remove it. Judged on that basis, a tool that discards everything it cannot resolve looks excellent.

The more useful question is not how much a verification pass removed. It is how much of what it removed was worth keeping.