Lead Lists Built With Provenance You Can Show a Regulator

Every record we deliver carries three extra columns: where it came from, the date it was obtained, and the lawful basis relied on. Under Art. 14 of the GDPR you have one month from obtaining personal data to notify the person, and that clock starts whether or not the vendor who sold you the file mentioned it. A list without provenance is not cheaper than one with it; it is the same list with the liability moved onto you quietly. What the work actually is: The target definition, not the extraction. We would rather hand you 400 accounts each attached to a named trigger than 40,000 filtered on job title and headcount, because the second list is what your competitors also bought and it converts accordingly. Triggers we build against: a role posted that implies the problem exists, a technology added or removed on the domain, a funding or filing event, a leadership change in the function that owns the budget, an expiring vendor contract visible in a public tender. Each trigger becomes a testable hypothesis about why this account should reply this quarter, and the ones that do not produce replies get retired rather than defended. Sources and why each: Cognism and Ocean.io for European records where the vendor has performed its own notification, Apollo for US coverage, Clay as the orchestration layer so the enrichment logic is inspectable rather than a black box, Sales Navigator for account and reporting-line structure. Company-level verification against Registro Imprese and VIES for Italian entities and Companies House for UK ones, because a target list containing dissolved companies is a signal about the whole file. Technographic filters from BuiltWith or Wappalyzer, checked against the live domain rather than trusted from a cached index. Verification and what we drop: A verification pass through MillionVerifier or ZeroBounce before delivery. Catch-all domains are segregated into their own file and flagged, never blended into the main set, because they inflate a deliverable count and land as bounces two weeks later in somebody else's ramp. Role addresses — info@, sales@, contact@ — are dropped rather than kept for coverage. What we refuse to deliver: Pattern-guessed addresses from first-name-dot-last-name permutation tools. Personal consumer mailboxes, including gmail.com and libero.it addresses attached to a business contact. Scraped consumer data. Any inference about health status, and for pharma and medtech clients any HCP list that does not originate from a licensed source where the physician received their own notice. Files a vendor describes as verified without naming the origin of each record. And we do not price per thousand records, because that price rewards exactly the behaviour that ruins a sending estate. What you receive: The target set as CSV and as a live Google Sheet, with provenance columns. The Clay table itself, so your team can rerun the build when the trigger fires again rather than commissioning us. Art. 14 notice text drafted to sit in a first message. A suppression file format that your sequencer and CRM will both accept. And a written note on the accounts we excluded and why, which is usually the most argued-over page. Not included: Sending. Sequence copy. Phone numbers and dialler data. Cleanup of your existing CRM beyond the records we deliver. Who this is not for: Anyone comparing quotes on cost per record. That number can always be beaten by someone guessing addresses.