How to clean a scraped email list
Scraped lists fail in predictable ways: the same person appears under several URLs, addresses are truncated at a visible-length cut, the @ is written as (at) or [at], and a share of the entries are deliberately planted trap addresses.
Clean in this order — extract and repair, dedupe, drop role accounts, kill dead domains — and expect the list to shrink by a third before you send to it.
1. Repair the extraction damage
Look for (at), [at], at , (dot), and addresses cut off mid-domain. Also check for HTML entities and trailing punctuation picked up from the page: jane@acme.com. and jane@acme.com are the same address, but only one parses.
2. Dedupe on the canonical address
Scraped data duplicates heavily — the same person on a team page, a press release and a conference listing. Normalise, then dedupe. Expect more duplicates than you think: it is common for a third of the raw rows to collapse.
3. Drop role and no-reply addresses
info@, contact@, support@, noreply@, postmaster@. They deliver, so they don't bounce — they just never engage, and a list that never engages trains providers to filter you.
4. Kill domains with no mail server
Scraped sources include domains that expired since the page was crawled. An MX check removes every address on a domain that cannot receive mail, and it costs nothing per address.
5. Accept what you cannot detect
Honeypots and spam traps look like ordinary addresses. A trap that has never opted in will report you, and no cleaning step removes it — only not sending to scraped lists in the first place does. If you send, send small and watch complaint rates, because Google's published guidance is a spam rate under 0.3% in Postmaster Tools.
The risk that isn't technical
In the US, CAN-SPAM requires accurate headers and a working opt-out, and it applies to scraped lists too. In the EU and UK, unsolicited commercial email to individuals generally needs consent, and a harvested address does not come with any. Cleaning makes a list technically sendable; it does not make it lawful to send. That distinction is worth more than any bounce-rate improvement.
Check your own list in about 30 seconds. Paste or drop a CSV — duplicates, typos, role addresses and dead domains, all handled in your browser. Free for the first 100 addresses, $19 once for unlimited.
Open the toolCommon questions
Does cleaning make a scraped list safe?
Technically safer, legally unchanged. Cleaning removes bounces and bad addresses; it does not create consent, and consent is what most jurisdictions actually require.
Why do some cleaned addresses still bounce?
Because MX records prove a domain accepts mail, not that a specific mailbox exists. Fabricated addresses on real domains pass every domain-level check.
Last updated 2026-10-23.