You drop in a hundred domains and a third of them come back empty. The contacts are there, you can see them with your own eyes, the parser cannot. Sound familiar?
We fixed that the radical way and gave Emails Finder a brain.
Contacts are now found by AI models we trained on millions of pages and addresses. More emails, more contact pages, and the first address in a row is the editorial desk or the ads department, not a faceless info@. Below are the numbers and the sites that used to return zero.
What changed
The parser used to wander around a site like a tourist without a map. Now it reads pages for what they mean. It does not care what language the site is in or what the contact page is called over there: Kontakt, Impressum, Contactez-nous or About us. It knows what it is looking at and goes where the people are.
The second change is the order. The first address is now the one most relevant for partnerships and business: the editorial desk, the ads department, a specific person, not a generic info@.
The numbers
The fairest test is the same sites before and after. We ran 20,000 domains that went through the tool before the update and after it. One in six came back with more contacts.
And here is the whole flow over the first week:
| Metric | Before | Now |
|---|---|---|
| Domains with a Valid Email | 54% | 76% |
| Domains with a Contact Page | 78% | 88% |
| Valid Addresses per Domain | 1.26 | 1.95 |
That is not a sample of a hundred sites. That is tens of thousands of domains every week.
About quality
When the number of addresses suddenly jumps, the first question is the right one: is it junk? No. The valid share of everything found was 85% and it is still 85%. One and a half times more addresses per domain, and just as many of them alive.
Every address is verified before it reaches the table. A green dot means valid, no surprises.
The first address in a row is the one to start with. It is valid in 87% of cases.
Examples
Now the fun part. Every one of these sites used to return zero: no address, no contact page. Look at them now:
- versicherungsbote.de: kontakt.html, carrying [email protected] and two named editor addresses.
- hardware-news.de: the contacts sit in the Impressum, where all of Germany keeps them by law.
- savvymom.ca: the advertiser page and [email protected]. The very person who decides about placements.
- tourismhamilton.com: the partners page, four valid addresses with names.
- commonwealthfund.org: staff contacts, by name.
- naitreetgrandir.com: a French site with the contacts buried four levels deep. Found.
- uopeople.edu: five valid addresses.
Why this matters for outreach
A letter to info@ is read by a receptionist or a spam filter. A letter to redaktion@ is read by someone who already has placements in their price list. Same pitch, two different fates.
The contact page talks too. An advertiser page means: they sell here. A partners page: they are open to working together. An Impressum: here is who answers for everything.
What you have to do
Nothing. It already works: in Emails Finder, in the API and in the Chrome extension. The price is the same: one credit per domain, and only for the ones where something was found.