Every setting that matters in the Work Email waterfall. On 10,000 contacts, the default configuration returns 2,712 emails. After three settings, the same cascade returns 5,012 valid ones, from the exact same inputs.
A Clay cascade of 10,000 randomly picked contacts. First name, last name, company name only. Two successive runs, same cascade, same inputs. An address counts as found only when the configuration keeps it, because a discarded address never gets sent.
This is the most common configuration mistake, and the most expensive one. Many tables start with a column that looks up a company domain from a company name, then feed that domain into the waterfall.
The problem is that a mistake at this step is invisible and final. If the returned domain does not belong to the right legal entity, every provider in the waterfall then works from a false premise. They fail in series, or worse, they find a valid address at the wrong company.
deloitte.com, deloitte.com.au, deloitte.co.uk, deloitte.es, deloitte.fr. Five domains for one brand, only one is right for this contact. Multiply that by every company in your table with a generic name, a shared name, or a legal name that differs from its trading name.
A valid email sent to the wrong domain is worse than no email at all. It lands in a real inbox, just not your prospect's. No bounce, no alert, no reply. Only a reply rate that quietly collapses.
Delete the domain lookup column and feed the waterfall what you already have: first name, last name, company name. Dropcontact's algorithms identify the real entity behind the name and resolve the matching domain, with no intermediate step. Measured across the 20,000 contacts of our benchmark: 1.0% inconsistent domains, the lowest of the fifteen tools tested, on a scale that runs up to 22.5%.
In a waterfall, the first provider that returns an address stops the sequence. Clay does not compare answers against each other, it keeps the first one. So provider order does not decide your budget, it decides which data enters your file.
Sorting providers from cheapest to most expensive is therefore a bad trade. A weak provider returns just as many addresses, with invisible errors: a domain that is not the right entity, or someone who has since left the company. The waterfall stops there and no later step corrects it. You paid once, for wrong data you will only discover in your reply rate.
Across the 20,000 contacts of our benchmark, Dropcontact comes out at 54.9% genuinely usable emails, measured after actually sending every address found, with 1.9% total loss between hard bounces and inconsistent domains.
There is a second reason, and it comes down to how the data is produced. A database built in the past ages fast: 25 to 33% of contacts change employer within a year, the stored address stays valid and the person is gone. A real-time algorithm recomputes on every call. But not all algorithmic solutions are equal: nearly every email finder handles two first names, two last names, or both, poorly, and the few algorithmic ones cannot resolve them. Maria del Carmen Garcia Lopez, Jan-Willem van der Berg, and most names in Portugal and Latin America come back empty unless the address already sits in a database. Dropcontact resolves them algorithmically, and the same algorithms run identically across North America, Europe and Latin America.
Open the full configuration of your Work Email column, drag Dropcontact to the top of the sequence, save. Use the Clay managed Dropcontact account: nothing to connect, 2 credits per enriched cell, charged only when the email is found, validated and verified. Those 2 credits also cover the company domain lookup and the enrichment when the contact has changed company.
The full protocol, the fifteen tools tested and the results by region are in our email finder benchmark on 20,000 contacts.
This is the setting that surprises people most, and the one that pays off most.
A catch-all domain accepts every message, including messages addressed to a mailbox that does not exist. A standard deliverability test therefore always answers yes, and proves nothing. Faced with that dead end, third-party validators all do the same thing: they flag the address as risky, and your waterfall discards it.
Those domains belong to a very recognizable category of companies: large accounts, tech leaders, hypergrowth startups. Your most strategic targets. On every run you delete your best prospects first, and nothing in your table tells you what you just lost.
And this is not a marginal slice of your file. Catch-all is what large organizations configure. So the addresses your validator silently deletes belong to the segment with the highest purchasing power, the longest contracts and the deepest decision chains. They are also the least solicited contacts on the market, precisely because everyone else's stack deletes them too. The result is that most teams compete over the same easy half of the market and quietly discard the half that pays best.
The Validation Provider field is optional. Leave it empty when Dropcontact sits first. Validation is already included in the 2 credits, catch-all domains included. A second validation layer only removes deliverable addresses, and you pay for it.
How the algorithms decide on a catch-all address, and what they return when they cannot, is detailed on our email verifier page.
Between 25% and 33% of the contacts in a database change employer within a year. Those addresses pass every test: the domain answers, the syntax is clean, the validator says yes. They are valid and completely useless.
That is the structural limit of any provider querying a database built in the past. And you lose twice: once on the dead email, once on the opportunity, because someone who just joined a new company is exactly the person to contact.
Dropcontact detects the change at enrichment time, updates the company and recalculates the work email on the new domain. Nothing to switch on: with the waterfall configured as described above, every enrichment benefits from it.
Outside Clay, the same detection can run continuously on a CRM and surface the contacts who just moved. See company change alerts and the list of native integrations.
Three minutes of configuration. Test on 500 rows, compare, then run your full table. On hard files, also try Dropcontact on its own: disable the other providers and compare. That is what the video shows, and it often wins on European files and on large accounts.
You will read that a waterfall gets you nine emails out of ten. Ask two questions before you believe it: on which file, and measured at which point.
A match rate inflates in two ways. Counting an address as found before deliverability is checked, and running the test on a pre-filtered list of large, well documented companies instead of a raw file. Both are legitimate ways to build a table. Neither produces a number you can compare to another one.
Our figures, with their sample: 50.1% on 10,000 contacts picked at random inside Clay, first name, last name and company name only. 54.9% genuinely usable emails across the 20,000 contacts of our benchmark, measured after actually sending every address found, with 1.9% total loss between hard bounces and inconsistent domains. Lower numbers than the ones you see quoted. Measured ones.
10,000 B2B contacts picked at random. Three input columns only: first name, last name, company name. No company domain provided.
May 2026. Both runs were made in the same Clay table, whose records carry a creation timestamp of 19 May 2026. The wider benchmark this page refers to was run in August and September 2025.
Two successive runs in the same Clay cascade. Run one used the default Work Email waterfall followed by a third-party validator. Run two applied the three settings above.
An address returned and kept by the configuration, as it would reach a sending tool. An address discarded by the validator is not counted, because it never gets sent. That choice is deliberate and it is the whole point of setting three.
Denis Cohen, co-founder and CEO of Dropcontact. Published 17 August 2026.
Dropcontact ran and funded this test. Dropcontact is one of the email providers available in Clay, so read it as a vendor test, not as an independent study. The wider study it refers to, our benchmark of 15 email finders on 20,000 contacts, is also run and funded by Dropcontact and publishes its full protocol.
One sample, one date, one waterfall configuration among many possible ones. Provider order, credit plan and file composition all move the result. Your own rate depends on your file, which is why these settings are worth three minutes on a 500 row test of your own.
Three settings, in this order: delete the column that guesses the domain, put Dropcontact first in the Work Email waterfall, and leave the Validation Provider field empty. On 10,000 contacts, these settings take the number of emails found from 2,712 to 5,012.
The one with the highest rate of genuinely usable emails, because the waterfall stops at the first valid result and no later step corrects it. Sorting from cheapest to most expensive puts the least reliable answer in charge: a wrong domain or a contact who has left the company still counts as a hit and still stops the cascade.
Because the third-party validator at the end of the waterfall cannot conclude on a catch-all domain, which accepts every message by design. It flags the address as risky out of caution, and the row is rejected. Those domains mostly belong to large companies.
No, not if the provider sitting first already validates what it returns. Dropcontact validation is included in the 2 credits per cell, catch-all included. A second layer costs credits and removes deliverable addresses.
2 credits per enriched cell with the Clay managed Dropcontact account, charged only when the email is found, validated and verified. Those 2 credits also cover the company domain lookup and the enrichment when the contact has changed company. Nothing to connect, nothing to configure.
First name, last name and company name are enough. The company domain improves the result further, but only if you hold it with certainty, from the source, and not if it comes from a domain enrichment tool. Do not add another solution in Clay to look it up: the domain lookup is already included in Dropcontact, and it performs better run directly by us.
Not on a raw B2B file, and any rate quoted without its sample is not comparable. Two things inflate a match rate: counting addresses before deliverability is checked, and running on a pre-filtered list of large, well documented companies. Our measured figures are 50.1% on 10,000 random contacts inside Clay, and 54.9% usable emails across the 20,000 contacts of our benchmark, measured after actually sending each address found. Before comparing two numbers, ask for the sample and the loss rate.
That is what you read everywhere online, repeated from one article to the next, and it is sound whenever nothing in your stack can actually verify a catch-all domain: sending to an unverified address is a fast way to burn a sending domain. The premise changes when the provider itself resolves catch-all addresses. Dropcontact returns a verified or invalid status on catch-all domains, so the address you keep has been checked rather than assumed. Deleting unverified addresses is right. Deleting verified ones because a second tool could not conclude is how you lose your largest accounts.
Dropcontact is one of the providers available in Clay, and we are the ones who filmed this demonstration. You have no reason to take our word for it.
The benchmark this page refers to is also run and funded by Dropcontact, which is among the fifteen tools tested. Its method is published in full: 20,000 real contacts, three input columns only, every email found actually sent, and independent double entry to verify each company and domain pairing. The protocol is reproducible.
And above all, these settings take three minutes to test on your own cascade. Then send the emails you get and look at what actually bounces, without running them through a third-party verifier that would discard your catch-all addresses. That is the only verification that counts.
Take a table you have already enriched. Apply the three settings to 500 rows. Compare the two columns.
Optimize my Clay setup with DropcontactClay managed Dropcontact account. Nothing to connect, 2 credits per domain and email found and verified.