Domain → Page → Contacts · the company page is the matching key
How to find employees of a company from its domain
The short answer
To find employees of a company from its domain, don't search people by the domain. Resolve the domain to the company's LinkedIn page first, check that it is the right company, then search the people whose current job is attached to that page. People are attached to company pages, not to websites, and any page can type any domain into its website field.
On this page11 sections
- Where are employees actually attached to a company?
- Why is a website a weak link between a company and its people?
- How many company pages can claim the same domain?
- What does a domain search decide without telling you?
- What happens when you search a company’s people by its domain?
- Why does the matching key matter so much?
- Why are domain to page and page to contacts two separate calls?
- How do you find employees of a company from its domain, step by step?
- How do you call Domain to LinkedIn and Employee Finder with the API?
- Sources and method
- Video transcript
Say you want to reach the people at a company. All you have is its domain. So you do what everyone does: domain in, contacts out. It looks simple, and it is how many tools let you find employees by domain. Except that’s not how the data is connected.
This guide shows how to find the employees of a company when all you have is its domain. It follows the reasoning of the video, with the same examples: where people are really attached to a company, why a website is a weak link, what a domain search decides without telling you, and why Domain to LinkedIn and Employee Finder are two separate calls.
Where are employees actually attached to a company?
To a company page, not to a website. Think about the contact you’re trying to find, and ask her where she works. She won’t say acme.com. She’ll say Acme. On LinkedIn, that answer is a link from her position to the Acme company page, and she chose that page herself, on her own profile.
The company page has a website field, and from that website you can extract a domain. So the real structure is a chain of four objects: person, company page, website, domain.
Person
LMLéa MartinVP Sales at Acme- Works for
Company page
AAcmeCompany page- Website
Website
https://www.acme.com- Extract
Domain
acme.com
The domain is an attribute of the company page. It is not the object that connects the person to the company, and that difference decides what a search by domain can find.
Contacts aren’t attached to websites. They’re attached to company pages.
Why is a website a weak link between a company and its people?
Because the website field of a LinkedIn company page is not verified: anyone can type anything. A medical practice can put its Doctolib booking page there. A design agency can put a Figma prototype, an architecture firm its Facebook page, a one-person company a Notion page. Each of those pages now points to someone else’s domain.
A completely unrelated company can even write microsoft.com. The page of Sabena, a Belgian airline that stopped flying in 2001, gives a Microsoft web page as its website. And the real Microsoft page doesn’t use microsoft.com at all: it lists news.microsoft.com.
- DA dental practiceMedical practiceWebsiteNot verifieddoctolib.fr/dentist/…Domaindoctolib.frits Doctolib booking page
- DA design agencyBranding and designWebsiteNot verifiedfigma.com/proto/…Domainfigma.coma Figma prototype
- AAn architecture firmArchitectsWebsiteNot verifiedfacebook.com/…Domainfacebook.comits Facebook page
- OA one-person companyFreelancerWebsiteNot verifiednotion.so/…Domainnotion.soa Notion page
- SSabenaBelgian airline, defunct since 2001WebsiteNot verifiedmicrosoft.com/belux/…Domainmicrosoft.coma Microsoft licensing page
- MMicrosoftThe real companyWebsiteIts ownnews.microsoft.comDomainmicrosoft.comits newsroom
So the chain has a strong link and a weak one. Between a person and a company page, the link is a position the person chose. Between a company page and a domain, it is a field anyone can fill. A domain search starts from the weak one.
How many company pages can claim the same domain?
Often more than one, and for well-known companies, many. On October 11, 2026, we checked the domains of 122 well-known companies, from large groups to widely used software: 79 of them were given as a website by at least ten other company pages.
Domain to LinkedIn shows it in its response. Ask it for figma.com and it returns the Figma page, then lists ten more pages that give figma.com as their website: Figma’s own product and community pages, design studios and agencies, a university design society. Pick a domain below to see what came back on November 10, 2026.
Domain
figma.comThe page it matches
FFigma10 other pages give the same domain
- Figma's own product and community pagesFigma WeaveFriends of Figma | AnkaraFigma Nepal
- Design studios and agencies6 pages
- A university design society1 page
Domain
microsoft.comThe page it matches
MMicrosoft10 other pages give the same domain
- Other pages named after MicrosoftMicrosoft for StartupsMicrosoft MechanicsMicrosoft India+5
- Companies Microsoft acquiredSkypeRevolution Analytics
Domain
notion.soThe page it matches
NNotion10 other pages give the same domain
- Companies whose website is a Notion pageSOCARWatchabackpackrPetFriends+5
- A student entrepreneurship club1 page
It isn’t only a big-brand problem. Any page whose website is a path on a platform gives the platform’s domain: a startup whose careers site is a Notion page gives notion.so, an event organizer that sells tickets online gives its ticketing platform. In the October tests, a ticketing company with 20 employees on LinkedIn shared its domain with two associations that list their ticket page as their website. Together, the associations had about 500 people on LinkedIn: the pages claiming the domain outnumbered the company.
What does a domain search decide without telling you?
Which company page the domain belongs to. If your workflow starts with microsoft.com → contacts, the tool has to make that identity decision somewhere in the middle. One call doesn’t skip the question. It answers it where you can’t see it, with one of three shortcuts:
- Merge every page that claims the domain. You get the company’s people mixed with everyone else’s.
- Pick one page and hope it’s the right one. When it isn’t, the whole list belongs to another company.
- Rely on the one domain it already knows. A company filed under doctolib.com comes back empty for doctolib.fr, without saying why.
- microsoft.com1 callMerge every page that claims the domainMicrosoftSabenaUMCThree companies mixed in one list
- backmarket.com1 callPick one page and hope it is the right oneRefurbisherA two-person refurbisher instead of Back Market
- doctolib.fr1 callRely on the one domain it already knowsNo one: the company is filed under doctolib.com
Different shortcuts. Same hidden risk.
What happens when you search a company’s people by its domain?
You can get real people at the wrong company. In the tests behind the video, run on October 11, 2026:
- Ask for Microsoft by microsoft.com, and one contact database returned employees of Sabena: airline pilots, flight attendants and station managers, listed as Microsoft.
- Ask for Doctolib by doctolib.fr, and dental surgeons and physiotherapists became Doctolib employees: the practices that list their Doctolib booking page as their website.
- Ask for Figma by figma.com, and communication agencies in Argentina and Abu Dhabi started appearing inside Figma.
Ask for
Microsoftby microsoft.com- Airline PilotSabena · Belgian airlineListed as Microsoft
- Flight AttendantSabena · Belgian airlineListed as Microsoft
- Station ManagerSabena · Belgian airlineListed as Microsoft
Ask for
Doctolibby doctolib.fr- Dental SurgeonDental practiceListed as Doctolib
- PhysiotherapistMedical practiceListed as Doctolib
- Co-founderDental centersListed as Doctolib
Ask for
Figmaby figma.com- Managing DirectorPR agency · ArgentinaListed as Figma
- Founding PartnerPR agency · ArgentinaListed as Figma
- Head of CommunicationsCommunications agency · Abu DhabiListed as Figma
The people are real The company they are joined to is not theirs
The contacts themselves can be perfectly real. The problem is not necessarily the data. It’s the system connecting the data.
Filtering on job titles doesn’t fix it, and it can make it worse. People who leave a company to found their own often stay filed under their former employer’s domain, now with a title of Founder or CEO. In the same tests, 16 of the first 25 decision makers returned for loreal.com had left L’Oréal. The page on former employees shows how to find them on purpose, and keep them apart.
Good data joined to the wrong company is still bad data.
Why does the matching key matter so much?
Because the key decides which people a search can reach before any data is read. The domain and the company page are both ways to say “this company”, but only one of them is what people are attached to.
- Definition Matching key also called join key
- The value two datasets share that you use to decide that two records belong together. To find the people at a company, the key that holds is the company page their jobs are attached to, not the company's website.
PreciselyA matching key is the attribute used to join records from two datasets, such as an email, a profile URL or a company page URL. In B2B contact data, a person's current job points to a LinkedIn company page, while the domain comes from the website field of that page, which whoever runs the page fills in. Many pages can give the same domain, so a domain is a weak key for people; the company page is the key their jobs are attached to.
The October tests measured it directly. Same contact database, same companies, same request: only the key changed. We asked for the people at 23 companies whose domain other pages claim, once by domain and once by LinkedIn company page, then checked a sample of the returned profiles by hand. Matched by domain, 21% of the people we checked had no link to the company. Matched by company page, 1.5%. On a list of 1,000 contacts drawn from domains like these, that is 210 people who don’t work there, against 15.
The volumes moved the other way. By domain, that database held 2,041 contacts for Microsoft, most of them from Sabena. By the Microsoft page, more than 50,000. Doctolib went from 60 to 3,333, and Figma from 62 to 4,073.
Same contact database · same companies · only the key changes
Same underlying data. Different relationship between the objects. Completely different result.
Why are domain to page and page to contacts two separate calls?
Because they are two different decisions. The first is about identity: what company does this domain actually represent? The second is about people: who actually works at this company?
- Definition Domain to LinkedIn also called domain to company page
- Turning a website domain into the LinkedIn company page it belongs to, before you look for any people. It answers one question: which company does this domain represent?
PreciselyThe BlitzAPI endpoint POST /v2/enrichment/domain-to-linkedin takes a domain and returns the LinkedIn company page and name it matches, or found: false, with the other company pages that give the same domain as their website listed in a separate field. Employee Finder, Waterfall ICP and Company Enrichment take the page URL as input.
- Domain to LinkedIn answers the first one. It returns the company page the domain matches, and lists in a separate field the other pages that claim the same domain.
- Employee Finder answers the second one. It takes that page and returns the people whose current job is attached to it: one explicit company identity, not everyone in the database who carries the same domain.
- Validate
- Override
- Choose an alternative
- Cache
- Principal Program ManagerMicrosoft
- Software EngineerMicrosoft
- Director, Enterprise SalesMicrosoft
And yes, that means two API calls. Separating them gives you control. You can validate the company match, override it, choose one of the other pages, cache it, and reuse it across workflows. When you already have the LinkedIn page, skip the first step entirely: then it’s one call, the right one.
Hiding two operations behind one endpoint doesn’t remove the complexity. It only hides where the mistake happened.
How do you find employees of a company from its domain, step by step?
Start from the domain, settle the company once, then work from the page. Five steps:
- Resolve the domain to its company page. Send the domain to Domain to LinkedIn. You get the page it matches, its name, and the other pages that give the same domain. If nothing maps to the domain, the response says so: don’t fall back to a people search by domain.
- Check the match before you search for people. Compare the page name with the company name you already have, in your CRM or your list. When they don’t match, look at the other pages in the response, or set the account aside for review. A check this simple keeps other companies’ people out of your list.
- Store the page with the account. Save the LinkedIn company URL and its company ID next to the domain in your CRM. From then on, every workflow starts from the page, and the first step never runs twice for the same company.
- Search the people attached to that page. Employee Finder returns a list of employees at a company, filtered by title, level, function or location. To get the one right person for a role, Waterfall ICP tries your titles in order, as the decision makers guide explains.
- Enrich from the same key. Company Enrichment returns the full company profile from the page, and each person you keep can go on to their work email, as the page on how to find emails by domain explains.
- page → contactsEmployee FinderThe people whose current job is on the page
- page → buyerWaterfall ICPThe best decision maker, by your title priorities
- page → profileCompany EnrichmentIndustry, size, headquarters, website
- page → recordYour CRMStored with the account, next to the domain
That is also why the APIs stay decomposed. Each block solves one data problem, and you decide how the blocks connect. The page is the key between them.
Show the answer key
- Search the employees of the page it returned. It came first, so it must be the company.The name tells you otherwise. You would get the dental studio’s people, listed under your account: the hidden mistake of a domain search, made in plain sight.
- Choose the Acme page from the other pages, check it, store it with the account, then search its employees.The decision is visible, so you can correct it before you pay for a single contact. Store the page and the next workflow starts from it.
- Skip the check and search people by acme.com directly.The question doesn’t go away. A domain search still has to pick a page for acme.com; it just does it where you can’t see it.
The same logic holds wherever a domain is all you have: a form fill, a website visit, a list of accounts. Settle the company first, then the people. For inbound leads, lead to account matching applies it in the other direction. The data resolution use case applies it to whole CRM imports, and the TAM guide shows how to build the list of companies in the first place.
For builders
Everything above runs on two endpoints. This part is for the people who will wire them up: the two calls, a real run, and what to check before you ship.
How do you call Domain to LinkedIn and Employee Finder with the API?
A domain in, the company LinkedIn page it matches out, with the other pages that give the same domain.
Send
domain- The website domain. A full URL works too: https://www.figma.com/pricing is read as figma.com.
Read
found- false when no company page maps to the domain. Treat it as an answer, not as an error to retry by domain.
company_linkedin_url, company_name- The page the domain matches, and its name: the two values to check against your CRM.
other[]- Other company pages that give the same domain as their website, each with its URL and name.
A company page in, the people whose current job is attached to it out, filtered and paged.
Send
company_linkedin_url- The page you validated at the first step.
job_title, job_description- Include and exclude lists. Square brackets match a whole title exactly: [Head of Sales].
job_level, job_function- Seniority (C-Team, VP, Director, Manager, Staff) and function, for the current job at that company.
country_code, continent, sales_region- Where the people live. WORLD means no geographic filter.
max_results, page- Up to 25 people per page. The response gives total_pages.
Read
results[]- The people: name, headline, location, profile URL, skills, education.
results[].experiences[]- Their job at that company, with company_linkedin_url and company_domain: the domain arrives as an attribute of the job, through the page.
page, total_pages- Where you are in the list.
Each company or person returned is one record; plans and rate limits are on the BlitzAPI pricing page.
Two calls, one real run
On November 10, 2026, we sent figma.com to Domain to LinkedIn:
curl -X POST https://api.blitz-api.ai/v2/enrichment/domain-to-linkedin \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"domain":"figma.com"}'The response, with three of the ten other pages kept:
{
"found": true,
"company_linkedin_url": "https://www.linkedin.com/company/figma",
"company_name": "Figma",
"other": [
{
"company_linkedin_url": "https://www.linkedin.com/company/figmaweave",
"company_name": "Figma Weave"
},
{
"company_linkedin_url": "https://www.linkedin.com/company/friends-of-figma-ankara",
"company_name": "Friends of Figma | Ankara"
},
{
"company_linkedin_url": "https://www.linkedin.com/company/figma-nepal",
"company_name": "Figma Nepal"
}
]
}The name matches, so the page goes to the second call. Here, the directors in sales and business development:
{
"company_linkedin_url": "https://www.linkedin.com/company/figma",
"job_level": ["Director"],
"job_function": ["Sales & Business Development"],
"max_results": 2
}Two people came back, out of 8 pages. One of them, trimmed to the job that matters and with the name replaced:
{
"full_name": "Sam Rivera",
"headline": "Director, Enterprise Sales | @Figma",
"experiences": [
{
"company_name": "Figma",
"job_title": "Director, Enterprise Sales",
"company_linkedin_url": "https://www.linkedin.com/company/figma",
"company_domain": "figma.com",
"job_is_current": true
}
]
}The chain from the video is in the data: the job points to the company page, and figma.com comes with it, as an attribute of the page.
Before you ship
- Compare company_name with the name you already hold before the second call. Route mismatches to review, and look in other for the right page.
- Store company_linkedin_url on the account. Resolve each domain once, then start every workflow from the page.
- Treat found: false as a result. Keep the account for review instead of searching people by domain.
- Searching people by a platform domain (a store builder, a site host, a ticketing service): it names thousands of sites, not one company.
- Filter people on the current job at the page (job_title, job_level, job_function), and page through with page until total_pages.
- When you already have the LinkedIn page, from a CRM field or a person’s experience, skip the first call.
Sources and method
- Written by
- The BlitzAPI team
- Published
- API calls run on
- Domain to LinkedIn URL, BlitzAPI API reference
- Employee Finder, BlitzAPI API reference
- Find People guide, BlitzAPI documentation
- Company Enrichment, BlitzAPI API reference
The matching tests behind the video and this guide ran on October 11, 2026, on the domains of large groups, widely used software companies and smaller platforms, in the contact databases of several data providers. The providers are not named. The figures on the matching key come from one of those databases, queried twice with the same companies: once by domain, once by LinkedIn company page, with a sample of the returned profiles checked by hand.
The BlitzAPI team ran the API calls shown in this guide on November 10, 2026. In the sample responses, the people's names are replaced; the companies, the field names and the structure are kept as returned.
Acme and Léa Martin are fictional, as in the video. The people in the diagrams are described by their real job titles and employers, without their names.
Video transcript
Read the full transcript5:06 · 15 chapters
0:00Find the people at a company from its domain
Say you want to reach the people at a company. All you have is its domain. So you do what everyone does. Domain in. Contacts out. Simple. It’s how almost every provider works. Except there’s one problem. That’s not how the data is connected.
0:20Contacts aren't attached to websites
Think about the contact you’re trying to find. Ask her where she works. She won’t say acme.com. She’ll say Acme. Contacts aren’t attached to websites. On LinkedIn, they’re attached to company pages. And she chose that page herself, right on her profile.
0:37Person, company page, website, domain
That company page has a website field. And from that website, you can extract a domain. So the real structure looks like this: Person → Company Page → Website → Domain. The domain is an attribute of the company page. It is not the object that connects the person to the company. And that difference matters.
0:58The website field: anyone can type anything
Because the website field on LinkedIn is not verified. Anyone can type anything. A medical practice can put its Doctolib booking page there. An agency can put a Figma prototype. An architect can put its Facebook page. A freelancer can use Notion. And a completely unrelated company can write microsoft.com. In fact, the real Microsoft page doesn’t even use microsoft.com. It uses news.microsoft.com.
1:25A strong link and a weak one
That chain has a strong link… and a weak one. A domain search starts from the weak one. So if your workflow starts with: microsoft.com → contacts the provider has to make an identity decision somewhere in the middle. Which company page does this domain actually belong to? One call doesn’t skip that question. It just answers it where you can’t see it. One system can merge every page claiming that domain. Another can choose one page and hope it picked the right one. Another can rely on one domain it already knows. Different shortcuts. Same hidden risk.
2:02The same matching test
We ran the same matching test across the other major data providers. The pattern kept coming back. On average: 21% unrelated contacts. About one contact in five. People attached to a company they didn’t actually work for.
2:19Ask for Microsoft, get Sabena employees
Ask for Microsoft… and one system can give you employees from Sabena, a Belgian airline that disappeared in 2001. Ask for Doctolib… and suddenly dentists and physiotherapists become Doctolib employees. Ask for Figma… and communication agencies start appearing inside Figma.
2:38Good data joined to the wrong company
The contacts themselves can be perfectly real. The problem is not necessarily the data. It’s the system connecting the data. Good data joined to the wrong company… is still bad data.
2:51The matching key: domain or company page
And this is why the matching key matters so much. Take the same contact database. Match through the domain: 21%. Match through the LinkedIn company page: 1.5%. On a list of 1,000 contacts, that’s 210 people who don’t work there… versus 15. Fourteen times fewer. Same underlying data. Different relationship between the objects. Completely different result.
3:20Two problems: which company, who works there
That’s why Blitz treats these as two different problems. First: What company does this domain actually represent? Domain → LinkedIn page. We return the best match… and expose the alternative pages claiming the same domain. Then comes another question: Who actually works at this company? LinkedIn page → contacts. Now we’re searching people attached to one explicit company identity. Not everyone somewhere in the database carrying the same domain.
3:49Why two API calls?
And yes… that means two API calls. "Why do I need two API calls?" "Every other provider lets me do it in one." We hear this a lot at Blitz. Because they are two different decisions.
4:02Validate, override and reuse the company match
Separating them gives you control. You can validate the company match. Override it. Choose an alternative. Cache it. Reuse it across workflows. Or skip the first step entirely when you already have the LinkedIn page. Then it’s one call. The right one.
4:18The company page is the key to everything else
And that page is the key to everything else: the people, the best decision-maker, the full company profile. That’s also why our APIs stay decomposed. Each block solves one data problem. And you decide how the blocks connect.
4:35One endpoint hides where the mistake happened
Because hiding two operations behind one endpoint doesn’t remove the complexity. It only hides where the mistake happened. At Blitz, we’d rather make that decision visible.
4:46Domain, page, contacts
A domain helps you find the company. The company page helps you find the people. Domain → Page → Contacts. No shortcut through the wrong object. At Blitz, we don’t give you what you want. We give you what works.