How to Find Decision-Makers and Their Emails from a Company URL

The hard part of prospecting from a domain is not finding names. It is choosing which three to five names matter and getting emails that land. A domain hands you everyone: the intern, the contractor, the VP who left in March. A scraper that returns 400 contacts at one account has not done your job. It has handed you a new one. The chain that matters runs domain to company to the right people to verified work emails, and every link narrows the set. Skip the narrowing and you ship to the wrong inboxes at the right company, which reads worse than silence. This is how to build it.

Step 1: Resolve the URL into a real company record

A URL is not a company. It is a string that usually points to one.

Before you look for a single person, the domain has to resolve into a firmographic record: legal entity, headcount, industry, and funding stage. The same domain can hide three different things. acme.io might be a 12-person seed-stage startup, a brand owned by a 4,000-person parent, or a parked redirect to acme.com. People searches run off the company identifier, so a shallow company record produces a wrong people list no matter how good your filters are. In Clay, you paste the domain into a table, run a company enrichment to pull the firmographics, and confirm the entity is the one you mean to sell into before any people search runs.

The detail that trips teams up is the parent-subsidiary split. If the domain resolves to a subsidiary but the buying authority sits at the parent, your search at the subsidiary surfaces managers who cannot sign anything. Pull the company hierarchy in this step so you know which entity owns the budget.

Step 2: Decide who counts as a decision-maker before you search

Most "decision-maker" searches fail because they never defined the term.

A decision-maker is not a seniority tier. It is a role relative to your specific deal. For a $2,000-a-month tool, the decision-maker might be a manager with a corporate card. For a six-figure platform, a committee decides, and the person with the title you would guess is often not the one who blocks or champions it. Filtering on "VP and above" returns the org's executives, not the people who decide on your category. Define the buying committee for your product first, then map titles onto each role.

The roles that matter most are the economic buyer and the champion, and they are rarely the same person. The economic buyer signs. The champion does the internal selling when you are not in the room. A list that captures the buyer and misses the champion stalls in legal. A list that captures the champion and misses the buyer never gets to legal. Map both before you search, and treat the end user and the blocker as context you enrich later.

Step 3: Run Find People filtered to those roles, not the whole org

Find People pulls contacts from the company; your filters decide whether it pulls the right ones.

In Clay, open the company table, click + Add, and search for Find People. The source searches off company attributes and lets you filter by job title (organizational level, function, or specific titles), experience, and location. The instinct is to fish: pull everyone and sort later. Resist it. Sourcing the list itself costs no Actions, but every contact you then enrich does, so a 400-person dragnet at one account spends downstream on people who will never open your email. Filter to the committee roles you defined in Step 2 before the search runs.

Two filtering rules do most of the work. First, filter on function plus organizational level, not exact titles. "Head of RevOps," "VP of Revenue Operations," and "Director, Revenue Ops" are the same role wearing three labels, and an exact-title filter silently drops two of them. Find People's title search already returns synonyms and similar titles unless you set it to match exactly, so lean on function plus level rather than typing one literal string. Second, scope your search tight enough that each account returns one to three plausible owners, ranked by how close they sit to your committee map, instead of a long tail you enrich blind.

The output is a people table where each row carries a full name, a professional profile, and the company domain inherited from the parent table. Those three fields are the inputs the email step needs. Without the domain in particular, the email finders downstream have nothing to construct or verify against, which is why resolving the company first (Step 1) is not optional.

Step 4: Confirm the right owner with an AI research column

A title is a guess about authority. The org chart is the answer.

Filters get you a short list of plausible owners. They cannot tell you which of two directors actually owns the function you sell into, or whether the "VP of Marketing" runs demand gen or brand. For that last-mile call, add a Use AI column set to Web research, or build a Claygent agent. Point it at the company's own site and the public web and have it name the owner for a specific function. This step separates a list of plausible contacts from a ranked list of the right ones.

Run that as a column against each account and you get a confidence-scored owner with a source you can audit. The "still here" flag catches the executive who left in March before you waste an email on them. Treat anything below high confidence as a contact to verify by hand, not to suppress. The point is not to automate the judgment away. It is to surface the judgment call so a human spends thirty seconds on the ten that are ambiguous instead of an hour on the hundred that are not.

Step 5: Find verified emails with a multi-provider waterfall

This is where single-finder approaches quietly fail. No one email provider covers your whole list, and the gaps are large.

Run your name-plus-domain rows through several email finders in sequence, stopping at the first that returns a verified result. The numbers below come from Clay's first-party test of work-email providers, and they explain why the sequence matters more than the brand.

The read is plain: Hunter returns the most accurate emails but reaches barely half a list, Findymail gives the best single balance, and each provider trades coverage for quality or the reverse. That is why you sequence them instead of picking one. In Clay, click Add enrichment, search "Work Email," and select the Work Email waterfall: it cascades across email-finding providers in sequence and stops at the first valid result. Here the two credit meters split. You spend Data Credits only on the provider that returns a match, so wider coverage does not mean proportionally higher data cost. Actions, the orchestration meter, still accrue per row you enrich. In Full configuration you can also toggle on Infer Email, a free first step that constructs an address from the name and domain. In Clay's internal testing, the default first.last@domain.com pattern returned a valid email about 31% of the time, with no provider call.

That match rate is the payoff of doing the chain in order. A clean company record, role-filtered people, and a stacked waterfall produce a list where nine in ten contacts resolve to a real, reachable person, not a list padded with addresses that bounce.

Step 6: Validate before send and gate the list

A found email is not a deliverable email. The two are different states, and only one is safe to send to.

Email finders return catch-all addresses, role inboxes, and stale addresses alongside the good ones. A catch-all looks valid because the mail server accepts everything, yet it may still bounce. A role inbox like info@ or sales@ reaches a room no one owns. Send to enough of either and your bounce rate crosses the threshold that damages domain reputation and lands the good emails in spam. So the last step before anything ships is a gate: every address clears validation and a role-check, or it does not go.

In the Work Email waterfall's Full configuration, the Validation section controls how strict the gate is. There is no default validation provider; you add one yourself, then toggle "Require validation success" so the waterfall accepts an email only when the provider confirms it. The Validation strategy sets your risk tolerance: Conservative for cold outreach where bounce rates affect sender reputation, Balanced when you want some catch-all coverage on a warm domain, or Aggressive when volume outweighs precision. Email results come back as Valid, Invalid, Catch-all, Unknown, or Role-based, so suppress Role-based and free-webmail addresses before they enter a personalized sequence. The list that comes out the far side is smaller than what the finders returned, and that is the correct outcome.

When the chain holds together, reps stop guessing. After consolidating its data stack and enriching the record around each buyer, one GTM team described the shift this way:

That is the difference the right contact set makes downstream: the rep opens with "who to call and what to say," not with a name and a guess.

Common failure modes

• Treating the domain as the company: A URL can resolve to a subsidiary, a parent, or a redirect. Pull the company hierarchy first so your people search runs against the entity that holds the budget, not a sibling brand.
• Filtering on seniority instead of role: "VP and above" returns executives, not the people who decide on your category. Define the buying committee for your product, then map titles to each role, and target the economic buyer and champion specifically.
• Pulling everyone and sorting later: Sourcing a list costs no Actions, but enriching each contact does. Scope Find People to one to three plausible owners per account and filter to committee roles before the search, not after.
• Trusting a single email finder: The best single provider reaches roughly half a list at top quality. Stacking providers in a waterfall reaches into the 90s and spends Data Credits only on the provider that wins each row.
• Sending the moment you have an address: Found is not deliverable. Gate every email through a validation provider and a role-check before it enters a sequence, especially on a new sending domain.