Mailing list scrubber

What a dirty list costs

Gmail, Outlook.com, Yahoo and the other large mailbox providers keep a running opinion of every sender. One of the things that weighs most is the share of your messages addressed to mailboxes that don’t exist. Each one comes back as a hard bounce, usually with the code 550 5.1.1. Keep producing them and you look like someone who bought or guessed addresses. Your next campaign then lands in spam for everybody on the list.

Most bad addresses were typed by real people in a hurry: gmial.com, hotmail.con, a trailing comma copied from a spreadsheet. Duplicates hurt in a different way. The same person gets the message twice and is more likely to press “Report spam”.

What happens to each row

1. Reading the file. The tool detects the encoding (UTF-8, UTF-16, otherwise Windows-1252), then the separator, from the first 60 lines. The email column is the one with the most cells containing @. You can correct all three by hand.
2. Harmless cleanup. Invisible characters, surrounding spaces, mailto:, a Name <address> wrapper, quotes, brackets and trailing punctuation come off. The domain is lowercased. The part before the @ stays as typed.
3. Shape of the address. Exactly one @, no space, 254 characters at most and 64 before the @, no dot at the start, at the end or doubled, a domain with at least one dot and an alphabetic ending.
4. Domain typos. The domain is compared with a list of about 80 widely used mailbox domains. The measure is edit distance, meaning the number of single-letter insertions, deletions, substitutions or swaps it takes to turn one into the other. One edit is tolerated for a five-letter name, two for longer ones if the first letter matches. Names shorter than five letters are never corrected. Endings that don’t exist, such as .con, .cmo or .comm, become .com. So do gmail,com and gmailcom.
5. Duplicates. Addresses are compared without regard to case. The first occurrence stays and later ones are dropped.
6. Disposable mailboxes. Addresses at about fifteen throwaway services such as Mailinator, YOPmail or Guerrilla Mail are rejected.

Typo fixes are only proposals. Nothing changes until you accept a fix, one at a time or all together.

Reading the statuses

Valid

Well formed, seen once, nothing to report.

To fix, then Fixed

A probable domain typo with the suggested replacement. “Likely typo” means a close match to a known provider. “Likely fix” is less certain, for example gmail.fr read as gmail.com.

Duplicate

Same address as an earlier row, and you’re given that row’s number.

Rejected

Malformed, disposable, or at a provider whose mail service has ended. The reason is shown.

Kept, with a note

A role address such as info@, sales@ or support@, which reaches a team and not one person. Also two rows that probably share a mailbox. Gmail ignores dots and anything after a +, so j.smith+news@gmail.com and jsmith@gmail.com deliver to the same place. These stay unless you choose to remove role addresses.

To watch

A provider that has announced the end of its mail service, or accented letters before the @, which many servers still refuse.

The clean file keeps your other columns, your header row and your separator, with the corrected address in place. It’s saved as UTF-8 with a byte-order mark so Excel opens it correctly. A second file lists the rejected rows with a reason column.

What a format check can’t know

This tool never contacts a mail server. It doesn’t look up the domain in DNS and doesn’t test whether the mailbox exists. A perfectly formed address at a real domain can still bounce because the person left the company last year.

No software can see a typo before the @, such as jonh.smith. Company and school domains aren’t compared with anything, so acme-crop.com passes. The list of closing providers only covers a handful of regional internet providers.

The real test is the first send. After it, export the hard bounces from your email platform and remove them for good. Our bounce message decoder explains the codes you’ll run into.