New AI Answers: get named when buyers ask ChatGPT, Perplexity and Gemini who to hire. Watch 77-second video

November 20, 2026 · LeadButton

Where Do Lead Generation Agencies Get Data? From 5 Sources, and 7 Questions Tell You Which One Yours Uses

Lead generation agencies pull data from five places: licensed B2B databases such as ZoomInfo and Apollo, contributor networks built from users’ inboxes, automated web crawling, scraped LinkedIn profiles, and their own manual research. Most agencies mix at least two of these, and few will tell you which.

Where the data comes from affects more than list quality. It determines what you owe a prospect under GDPR, whether a CAN-SPAM violation counts as an aggravated one, and whether a court order can make the data disappear after you have paid for it. This guide is part of our outbound compliance hub. We checked every vendor statement and rule below on its own page on 19 November 2026. None of it is legal advice.

Key takeaways

  • ZoomInfo says it scans more than 28 million sites and domains a day and relies on more than 200,000 professionals who share contacts in exchange for free access. Apollo says its contributor network has more than 2 million contributors.
  • LinkedIn sued the scraping API Proxycurl in January 2025. Proxycurl had shut down by July 2025, and a permanent injunction ordered it to delete LinkedIn data obtained through unauthorized means and to send the injunction to its customers.
  • CAN-SPAM treats sending to addresses harvested by automated means from a site that promises not to share them as an aggravated violation.
  • GDPR Article 14 says you must tell a prospect where their data came from, no later than your first message.
  • California charges registered data brokers a $6,000 fee, plus $200 a day for each deletion request they fail to honor.

Where do lead generation agencies get their contact data?

Nearly always from some mix of the five sources in the table below. A licensed database provides the base list, and research or verification tools fill in the gaps. The risky slice is whatever comes from scraping, since that data can be inaccurate and contractually tainted at the same time.

Source How the data is built Main risk to you What to ask
Licensed B2B database (ZoomInfo, Apollo, Cognism and similar) Vendor combines crawling, partners, contributor data and research, then licenses seats License terms may bar use by or for third parties; accuracy varies by segment Whose license is it, and does it permit work on my behalf?
Contributor network Users connect mailboxes, CRMs or contact files; the vendor extracts business contact details The people in the list never dealt with the vendor, which matters under GDPR and California law Which vendor, and what notice does it give the people in its database?
Automated web crawling Bots read company sites, press releases, filings and job posts Harvesting addresses from sites that promise not to share them is an aggravated CAN-SPAM violation Do you scrape any site yourselves, and which ones?
Scraped LinkedIn profiles Browser extensions, fake accounts or scraping APIs copy profile data LinkedIn sued one scraper in 2025 and won a deletion order; tools can disappear mid-contract Does any part of my list come from a LinkedIn extension or scraping API?
Manual research People build lists by hand from public sources and verify emails Slow and costly, so it is often claimed more than done What share of my list is hand-built, and can I see a sample with sources?
Source types as described by vendors on their own sites. Checked 19 November 2026.

How do ZoomInfo and Apollo build their databases?

Both describe a blend of machine collection and a contributor network. ZoomInfo’s data sources page lists four inputs: automated review of public information on “over 28 million sites and domains every day,” data partners, a contributory community of “more than 200,000 business professionals,” and an in-house research team. Apollo says it relies on a “data network of over 2 million contributors.”

Most buyers overlook the contributor model. According to Apollo’s help center article on its contributor network, once a customer connects a mailbox, automated algorithms may pull out business contact information, which Apollo says is “most commonly used as part of our verification process.” Apollo says it adds new contact data only after checking it “against multiple data sources,” does not link added data to the customer who supplied it, and does not read email content. You cannot opt out on your own. Turning off the sharing takes a custom plan arranged through Apollo’s sales team.

ZoomInfo’s page states that its tools “do not attempt to bypass password restrictions or CAPTCHAs” and that it does not gather personal browsing history or data on people acting purely in a personal capacity. Neither page gives a total record count or an accuracy rate, so any accuracy number an agency quotes for these databases is the agency’s own claim.

Is it legal for an agency to scrape LinkedIn for leads?

Scraping LinkedIn violates LinkedIn’s user agreement, and LinkedIn takes that to court. For a buyer, the risk is less about criminal charges and more practical: the tool your agency depends on can be shut down, and the data can be ordered deleted.

In January 2025, LinkedIn brought a federal suit against Proxycurl, a LinkedIn data API. On 28 July 2025 LinkedIn said it had resolved the case, which it described as being about “unauthorized data scraping, fraud, and trademark misuse.” Social Media Today reported that under the permanent injunction Proxycurl must delete LinkedIn data obtained through unauthorized means, stop accessing LinkedIn, and send the injunction to its customers. Proxycurl had already closed earlier in the month.

LinkedIn does not always need a lawsuit. Around 6 March 2025, the company pages of Apollo and Seamless.ai, which both sold Chrome extensions used to extract data from LinkedIn, vanished from the platform, according to MarTech. On 10 March 2025 Apollo’s CEO said the restriction had not disrupted Apollo’s services. On the account side, our guide to whether LinkedIn will ban you for automation explains what the user agreement forbids.

Does the data source matter under CAN-SPAM?

Yes, in one particular way. CAN-SPAM does not require consent for B2B email, but an ordinary violation becomes an aggravated one when the address was harvested by a bot from a site that says it will not share addresses, or was produced by guessing combinations of names and letters.

You will find the rule at 15 U.S.C. 7704(b)(1)(A). It kicks in when the sender knew, or should have known from the circumstances, that the address was obtained “using an automated means from an Internet website or proprietary online service operated by another person” whose notice said it would not transfer addresses. The second part covers “dictionary attacks,” which sits uncomfortably close to how some agencies guess first.last@ formats and send without checking. The basic CAN-SPAM rules that apply to every send are in our guide on whether cold email is legal in the US.

What do GDPR and California law require you to know about the source?

GDPR requires you to tell each prospect where their data came from, at the latest in your first message. In California, companies selling data about people they have no direct relationship with must register as data brokers and handle deletion requests on a set schedule.

GDPR Article 14 applies any time personal data was not collected from the person directly, which covers every bought or scraped list. Under Article 14(2)(f) you must disclose “from which source the personal data originate, and if applicable, whether it came from publicly accessible sources.” Article 14(3)(b) sets the timing: if the data is used to contact the person, “at the latest at the time of the first communication.” If an agency cannot name its source, it cannot help you meet that obligation. Our guide to whether cold email is legal under GDPR covers which European markets allow B2B cold email in the first place.

California’s rules apply to the vendor, but they still shape your list. According to CalPrivacy, a data broker is a business that knowingly collects and sells “the personal information of a consumer with whom the business does not have a direct relationship.” Registered brokers pay a $6,000 annual fee. Since 1 August 2026 they have had to download deletion lists from the state’s DROP platform at least every 45 days, and starting 1 January 2027 they must process those requests within 30 days, per CalPrivacy’s DROP guidance.

Rule What it requires Number to know
GDPR Art. 14(2)(f) and (3)(b) Tell the prospect the source of their data No later than the first message
CAN-SPAM 15 U.S.C. 7704(b)(1)(A) Harvested or dictionary-generated addresses aggravate a violation Applies on knowledge “fairly implied from objective circumstances”
California data broker registration Register each year the definition is met $6,000 fee; $200 a day if late
California DROP deletions Download and honor consumer deletion lists Every 45 days since 1 Aug 2026; 30-day processing from 1 Jan 2027; $200 a day per unhonored request
From the statute and regulator pages linked above. Checked 19 November 2026.

Our guide to CCPA rules for prospect data covers whether the CCPA itself applies to your company and how work emails came under it in 2023.

How do agencies describe their own data sources?

Mostly in vague terms. Of the three pages we looked at, only Belkins names the outside tools behind its lists, and none of them says what share of a list comes from each source.

Naming a tool does not name a source. “We use Apollo” tells you which database it is. It does not tell you whose license the agency works under, whether exports are permitted for client work, or what Apollo told the people in its database.

What should you ask an agency about its data before you sign?

Ask these seven questions and get the answers written into the contract, not just said on the sales call. If the agency cannot answer the first three clearly, you will not be able to meet your own notice duties in Europe or judge your exposure in the US.

Question A good answer A red flag
1. Which databases and tools build my list? Named vendors, by name “Proprietary sources”
2. Whose license covers the data? The agency’s, with terms that allow client work, or yours Seats shared across clients
3. Does any of it come from LinkedIn scraping? No, or a clear description of how and the risk Silence, or “everyone does it”
4. How are emails verified before sending? A named verification step and a bounce threshold Guessed formats sent unverified
5. What do you tell prospects about the source? Wording you can review for EU and UK sends Nothing
6. Who owns the list when the contract ends? A written answer either way Undefined
7. How do opt-outs and deletion requests reach suppression? A shared suppression list, updated per send Handled per campaign, by hand
Questions drawn from the rules above.

Who pays if the answers prove wrong is a matter for the contract. Our guide on who is liable for an agency’s cold emails explains why the client usually shares the exposure regardless of what the contract says.

Frequently asked questions

Where do lead generation agencies get email addresses?

Mainly from licensed B2B databases like ZoomInfo, Apollo and Cognism, supplemented by email-finding tools, web research and, at some agencies, LinkedIn scraping. Find out which before you sign.

Is buying a lead list legal?

In the US, generally yes for B2B email, provided each send complies with CAN-SPAM. In the EU and UK you still need a lawful basis, and you must tell each person where their data came from by your first message.

Can an agency scrape LinkedIn for my campaign?

Doing so breaks LinkedIn’s user agreement. LinkedIn sued the scraping API Proxycurl and won a permanent injunction requiring the scraped data to be deleted.

Do I have to tell prospects where I got their data?

Yes under GDPR Article 14, and no later than your first communication. US federal law has no general equivalent for B2B email.

Is ZoomInfo or Apollo data GDPR compliant?

Both publish pages on privacy and data sources, but whether you are compliant depends on how you use the data. The duty to notify each prospect stays with whoever sends the email.

How LeadButton handles this

LeadButton publishes its prices. Launch costs $1,500 a month for email, Growth costs $3,500 a month and adds LinkedIn, and Scale is priced on request. Whichever agency you pick, us included, ask it the seven questions above on the first call and get the answers in the contract.

The plans are laid out side by side on the LeadButton pricing page. We are not lawyers, and this is not legal advice.

Sources

Leave a Reply

Back home