A B2B email finder answers two questions, and most tools only attempt one
Ask a finder for the email of a head of operations at a mid-size logistics company and it returns an address. Two very different questions were hiding inside that request: does this person still hold that role, and does this mailbox accept mail today. An address can be right about the first and wrong about the second, which is how a finder produces rows that look correct and bounce anyway.
Three ways a finder produces an address
| Method | What it establishes | Typical failure |
|---|---|---|
Pattern guess — first.last@, f.last@, first@ | That the domain exists and the pattern is common | The pattern may route to a different person, or to an accept-all server that confirms nothing |
| Stored database lookup | That the address existed when the record was collected | The record ages: people change jobs and mailboxes get deleted |
| Live source scan — website, Telegram, Discord | That the address is published somewhere right now | Published does not mean monitored; a contact page can list a dead alias for years |
None of the three is dishonest on its own. The problem is the report that follows: a pattern guess and a live, handshaken mailbox arrive in the same column, formatted the same way, and get treated as equally true.
Why guessed addresses bounce later
- 23% of business contacts change jobs every year (ZoomInfo, 2025) — the person part of the row goes stale on a clock.
- 40% of email addresses are dead within two years (NeverBounce) — the mailbox part does too, faster than most teams re-buy a list.
- Freshly sourced data outperforms stored records by 42% (Harvard Business Review, 2024) — which is a statement about age, not about suppliers.
The last row of the first table is the one that hurts: an accept-all server says yes to anything@, so a check against it cannot distinguish a real person from an invented string. Tools that write that answer down as valid have moved an unknown into your bounce rate.
What to ask before trusting a finder
- Is there a verification state per row? Not a badge on the export — a per-address verdict from the same call that found it.
- Is the source of each row visible? A row found on a company site and a row generated from a pattern are not the same asset.
- How is a catch-all domain reported? If the answer is "valid", the tool is guessing.
- When was it checked? A verification result is a timestamp, not a permanent property.
How TAPAC returns it
One request searches by industry, job title, company size and location, scans the sources you allow, and probes the mailbox over SMTP inside the same call, so the row arrives with its verdict attached:
curl -X POST https://tapacapi.com/v1/contacts/search \
-H "Authorization: Bearer ***" \
-H "Content-Type: application/json" \
-d '{"industry":"logistics","job_titles":["head of operations"],
"company_size":"51-200","location":"poland","source":"website","limit":40}'
Each result carries the person, the company, the title, the source the address was found in, and the verification state of that mailbox. Accept-all domains come back unknown; a rejected mailbox comes back invalid. Addresses can also be checked on their own through POST /v1/contacts/verify.
From an agent it is one tool call — npx -y @tapacapi/mcp or the hosted endpoint https://tapacapi.com/mcp, exposing tapac_find_contacts and tapac_status. Without a key the server returns onboarding instead of an error. 100 searches are free, then $0.10–0.50 per verified contact, pay-per-use: no seat licence, no separate enrichment line.
The honest limit. Verification proves the mailbox exists; it says nothing about whether anyone wants your offer. And SMTP probing is rate-limited per domain, so large volumes need pacing — the server handles that, not the caller.
Where to go next
- Email list cleaning · B2B email finder · B2B lead list
- Catch-all verification · cold email bounce rate · what an SMTP check proves
- Install the TAPAC MCP server — npx, uvx or the hosted endpoint, with the JSON block for Claude, Cursor, Codex and Windsurf.
- Compare TAPAC with Apollo, ZoomInfo, Hunter and Lusha · Pricing · Get a free API key