What do email headers tell you?
The From address you see in the inbox is just text the sender wrote. The headers are harder to fake because each server that handles the message adds its own lines, and the last one, your own provider, also writes down the result of its authentication checks.
How to open the full headers
- Gmail: open the message, ⋮ menu › Show original.
- Outlook on the web: ⋯ menu › View › View message source.
- Outlook desktop: open the message › File › Properties › Internet headers.
- Thunderbird: View › Message Source, or Ctrl+U.
Work on a copy (save the .eml file) and do not click anything inside the message while you investigate.
The headers that matter, in order
Return-Path: <bounce@mailer.cheap-hosting.example>
Authentication-Results: mx.company.example;
spf=pass smtp.mailfrom=mailer.cheap-hosting.example;
dkim=pass header.d=cheap-hosting.example;
dmarc=fail header.from=bank.example
From: "Bank Security" <security@bank.example>
Reply-To: <support-bank@freemail.example>
Received: from mailer.cheap-hosting.example (203.0.113.45) by mx.company.example
Message-ID: <a1b2c3@mailer.cheap-hosting.example>| Header | What to check |
|---|---|
| From | The address the user sees. Easy to forge; never trust it alone. |
| Return-Path | Where bounces go, also called the envelope sender. A different domain from From is common for mailing services, but worth noting. |
| Reply-To | Where your reply would go. A free mail address here, under a corporate From, is a strong phishing sign. |
| Authentication-Results | What the receiving server checked: spf, dkim and dmarc, each with pass, fail, softfail or none. |
| Received | One line per server. Read from the bottom (the origin) to the top (your server). Look at the first external IP. |
| Message-ID | Generated by the sending system. Its domain often reveals the real sending platform. |
SPF, DKIM and DMARC without the jargon
| Check | Question it answers |
|---|---|
| SPF | Is this server allowed to send mail for the envelope sender's domain (Return-Path)? |
| DKIM | Was the message signed by the domain in header.d, and has it not changed since? |
| DMARC | Do SPF or DKIM pass for the same domain the user sees in From, and what does that domain ask receivers to do if not? |
The trap is that SPF and DKIM can pass for the attacker's own domain. DMARC is the check that ties the result to the visible From, so dmarc=fail on a well-known brand is the line to look for first.
A quick checklist
- 1Does the From domain match the organisation it claims to be? Look for look-alike domains.
- 2Does Reply-To point somewhere else, especially a free mail service?
- 3What do SPF, DKIM and DMARC say, and for which domain does each one pass?
- 4Which IP address sent it (the first external Received line), and who owns it?
- 5Do the links in the body point to the domain the message claims to come from?
Header analysers such as Google Admin Toolbox Messageheader or MXToolbox lay out the Received chain and the delays between hops, which saves time on long headers.
