The part of an email most people see, the sender name, subject and body, is the part the sender controls completely. The header, which most mail clients hide behind an option like “Show original” or “View message source”, is a mix of fields the sender wrote and fields that each mail server along the route added. Learning to tell those two apart is the whole skill of header analysis.
The fields the sender controls
Several header fields are written by whoever composes the message and are trivial to fake:
- From: the display name and address shown to the reader.
- Reply-To: where replies will go. If it differs from From, replies silently go somewhere else.
- Subject, Date and To: all set by the sender’s software.
- Message-ID: a supposedly unique identifier, normally generated by the sending system and often containing its domain.
These fields describe what the sender wants you to believe. They are useful mainly when they contradict each other or contradict the fields below.
The fields the servers add
Received lines are added by each server that handles the message. Each new server adds its line at the top, so the header reads like a stack: the newest hop is first and the oldest is last. To trace a message from origin to inbox, read the Received lines from the bottom up.
Each line typically says which host the message was received from (with its IP address), which server received it, and a timestamp. The important subtlety is trust. Lines added by your own organisation’s mail servers are reliable. Lines further down, added before the message reached your infrastructure, could have been forged by the sender to create a false trail. In practice, the most valuable line is the one where your own edge server records the IP address that connected to it, because the sender cannot fake that.
Return-Path records the envelope sender used during delivery. This is the address bounces go to and the domain SPF checks.
Authentication-Results is added by the receiving server and summarises the SPF, DKIM and DMARC outcomes, including which domain each check evaluated. It is often the fastest single line to read.
DKIM-Signature carries the signing domain and selector. ARC headers (Authenticated Received Chain) may appear when intermediaries like mailing lists or forwarding services preserve earlier authentication results. Many platforms also add their own fields with spam scores and filtering verdicts; these vary by vendor.
Walkthrough: the urgent vendor payment
An accounts executive at a manufacturing company in Coimbatore reports an email. The display shows “Ravi Kumar, Accounts” from a familiar supplier domain, asking for an updated bank account to be used for this month’s payment of ₹18.4 lakh. Here is how an analyst might reason through the header.
Step 1: compare From, Reply-To and Return-Path. From shows the supplier’s real domain. Reply-To points to a free webmail address with a similar name. Return-Path is on an unrelated domain. Three identities where a normal message has one or two that share a domain is an immediate concern.
Step 2: read Authentication-Results. SPF passed, but for the unrelated Return-Path domain, not the supplier. No DKIM signature from the supplier’s domain is present. DMARC for the supplier’s domain failed. The supplier publishes only a none policy, which is why the message was delivered instead of rejected.
Step 3: trace the Received chain. Reading upwards from the bottom, the earliest hops show a hosting provider’s server, not the supplier’s usual mail platform. The analyst checks a few genuine past emails from the same supplier, which consistently arrived via a major cloud mail service. The difference is clear.
Step 4: check the timeline. The Date field says 10:02 IST, but the first Received timestamp written by the company’s own gateway is 03:41 IST. Small gaps are normal; large gaps or a Date field later than the actual receipt can suggest the field was set manually. Remember to normalise timezones: a header may show +0000 on one line and +0530 on another.
Step 5: conclude. This is a spoofed message relying on the supplier’s weak DMARC policy, with replies diverted to an attacker-controlled mailbox. The analyst blocks the sending infrastructure and the reply address, searches for other recipients of similar messages, and alerts the finance team to verify any bank change by phone using a number already on file.
When the header looks perfect
A harder variant: every check passes. SPF, DKIM and DMARC all succeed for the real supplier domain, the Received chain shows the supplier’s normal mail platform, and From and Reply-To match. This is often a sign the supplier’s real mailbox has been compromised. The header is genuine because the message genuinely came from that account.
Here the header proves authenticity but not intent, and the analyst turns to other signals. Did the supplier’s mailbox suddenly start sending from a new country or new device? Is the request out of pattern, like a bank change mid-cycle? Are there odd inbox rules hinted at by the conversation, such as replies arriving faster than a human would type? Contacting the supplier through a known channel settles it.
A short checklist worth internalising
- Read Received lines bottom-up, and trust only those added by systems you control.
- Check which domain each authentication result applies to, not just whether it says pass.
- Look for disagreement between From, Reply-To and Return-Path.
- Compare against a known-good message from the same sender where possible.
- Treat the header as evidence about origin, not about whether the request is legitimate.
Most of the time, a few minutes with the header turns a vague “this feels off” into a specific, defensible conclusion, and tells you exactly what to block.
Written by the RP Strativa Team. The company, people, amounts and times in the example are illustrative.
Leave a Reply