How mailfloss measures email formats, verification and catch-all status

Where the data comes from

When a customer verifies a list, mailfloss checks each address with its domain's mail server and records the domain, the day and the kind of answer. Each night mailfloss rebuilds the reports from the last 90 complete days. A domain keeps its latest report, dated on its page, until a newer 90-day window produces one.

How email formats are classified

mailfloss sorts the part of each address before the @ into a fixed list of formats, such as first.last, flast or firstlast, using dictionaries of given and family names from many languages. Addresses it cannot sort confidently are reported as unclassified or as two unrecognised words, not guessed. The classification is reproducible but not perfect: names that are both a given and a family name, and some romanised names, can be misread.

A format's share is its fraction of the distinct addresses mailfloss verified, so an address verified many times still counts once. Distinct-address counts are lower bounds. The 95% range shows where the share can sit given the number of addresses behind it.

How verification behaviour is measured

Accepted, rejected and no-clear-answer shares come from the answers mail servers gave when mailfloss verified the domain's addresses. SMTP responses are reduced to their status code, enhanced status code and a fixed category, so no server text, address or IP address is ever stored or shown.

Pages list responses about the mailbox only: accepted, mailbox does not exist, inbox full, account disabled, mailbox receiving mail too fast, and no mail server.

How a domain is compared with its peers

A report compares the domain with the reported domains that route mail through the same provider, itself included: its accepted share, its leading listed format's share and its rejected share against the median of the group. The comparison appears only when the group is large enough for a median to mean something.

How catch-all status is decided

mailfloss checks whether a domain accepts mail for addresses that do not exist. The page shows one of four states, with the date of the latest check that gave an answer:

• Mostly accepts any address (catch-all).
• Mostly rejects unknown mailboxes.
• Inconclusive: recent checks disagreed, or there is no clear answer yet.
• Not checked recently: the latest result is 90 or more days old.

How mail providers and DNS records are identified

mailfloss reads each domain's public MX, SPF and DMARC records. The mail provider shown is the one the published MX record points to. That can be a security gateway, such as Proofpoint or Mimecast, that filters mail before it reaches the mailbox provider; where mailfloss also knows the mailbox provider, the page lists it separately.

Rounding and what is left out

• A domain gets a page only when enough of its addresses were verified in the last 90 days.
• Uncommon formats, outcomes and responses are combined or left out, and every count is rounded.
• A share is shown as a range whenever rounding or a combined remainder could move it by more than a point, and never as exactly 0% or 100% unless the data proves it.

Free-mail, disposable and education domains never get a page. The current set is a first batch of company domains; more follow once search engines have indexed it.

Removing a page

Whoever controls a domain can remove its page, with no wait for a person. Prove control in one of two ways: open a one-time link we email to one of the domain's admin addresses (admin@, administrator@, webmaster@, hostmaster@ or postmaster@), or add a DNS record we give you. The page comes down within minutes of the proof and stays down through every rebuild. A request without proof removes nothing, because the pages hold aggregates only. Remove a domain page.

Every domain page links to this at the bottom. Browse the company email domain reports to find a page.