How to read email headers and identify the sending service
Your business name can appear on an email sent by a mailbox, a CRM or a billing platform. The received headers help you distinguish those routes. This walkthrough shows which fields to read, how to compare two messages and what still needs confirmation from your administrator or software supplier.
- How can you tell which service sent an email?
- Start with the original received message. Identify the receiving system’s trusted Received and Authentication-Results fields, then compare the sending host, envelope domain and DKIM signing domain with the application’s sending record. The visible From address alone does not identify the sending service. Headers can identify an infrastructure route, but a CRM using your normal mailbox may need application logs to distinguish it from a manual send.
1. Obtain the original received headers
Use the message in the recipient’s mailbox, including junk where relevant. A copy in your Sent folder does not contain all the fields added during delivery. A normal inline forward creates a new outer message, so its headers may describe the forwarding journey instead.
In new Outlook, open the message’s More actions menu, then select View and View message details. In classic Outlook for Windows, open the message in its own window and use File, then Properties, to find Internet headers. Microsoft provides the current Outlook header instructions (opens in a new tab).
In Gmail in a browser, open the message, use the More menu beside Reply and choose Show original. Google explains this in its full-header guide (opens in a new tab). The location and labels can differ between clients, so use your provider’s instructions for the version you have.
Keep an untouched original securely. Headers can contain addresses, routing details and message identifiers. Share them only with authorised people through an agreed channel. If a recipient cannot provide them, start with the sending event or bounce and record that receiving-side evidence is missing.
2. Establish which system added the evidence
Ask the receiving administrator which gateway and authentication service handled the message. Use the fields that this trusted system added or accepted under its configured trust rules. A familiar server name or a line saying pass is not sufficient on its own, because header text can be forged.
The Authentication-Results standard (opens in a new tab) explains this trust problem. Do not simply pick whichever result looks most favourable when several Authentication-Results fields are present.
Received fields record successive handovers, with newer entries normally added above older ones. You can reconstruct the reported journey from bottom to top, but begin the investigation at your known receiving infrastructure and establish how far back the evidence is trustworthy. Lower entries are not automatically reliable. The SMTP trace specification (opens in a new tab) describes how these records are added.
3. Read an annotated example
This is a synthetic teaching example created for this guide. All addresses and hosts use reserved example domains, and the IP address is reserved for documentation. These excerpts are not a real delivery, a client message or a cryptographically verified test. Other fields, including the signature itself, are omitted.
The identity the recipient sees
From: Example Accounts <accounts@sender.example> Reply-To: finance@sender.exampleFrom presents the message’s author identity. Reply-To specifies where a normal reply should go. Neither line says which server connected to the recipient or whether the message came from Outlook, a CRM or an invoice tool.
The envelope return path
Return-Path: <bounce-472@mailer.example>This is the envelope return address recorded at final delivery, normally used for delivery-status messages. It differs from the visible From domain here. That can be a supported platform arrangement, rather than evidence of a fault by itself. Forwarding can also change the envelope address.
The handover observed by the receiver
Received: from outbound.mailer.example (outbound.mailer.example [192.0.2.25]) by mx.recipient.example with ESMTPS id DEMO472 for <buyer@recipient.example>; Tue, 03 Nov 2026 10:15:00 +0000Assuming this is the record added by the trusted receiving gateway, it reports that gateway accepting a connection from the shown host and IP at the stated time. Match that observation to the supplier’s infrastructure and logs. A branded host name alone does not establish which customer application initiated the message.
The receiver’s authentication results
Authentication-Results: mx.recipient.example; spf=pass smtp.mailfrom=mailer.example; dkim=pass header.d=sender.example header.s=demo; dmarc=pass header.from=sender.exampleThe synthetic receiver reports SPF passing for mailer.example and DKIM passing with sender.example as the signing domain. SPF’s domain differs from the From domain, but the reported DKIM domain matches it. That aligned DKIM pass can satisfy DMARC. None of these results establishes inbox placement or that the content is safe.
A reference for matching records
Message-ID: <invoice-demo-472@mailer.example> Date: Tue, 03 Nov 2026 10:14:58 +0000Use the full Message-ID alongside the recipient and timestamp when matching application and mail-system records. The identifier’s domain is a clue, not proof of the originating software. The Date field is supplied with the message and should be checked against trace timestamps rather than treated as an independent delivery clock.
The useful conclusion from this example is specific: the receiver reports a handover from the illustrated sending infrastructure and an aligned DKIM pass. To name the invoice application, we would still match the message to its sending log. To establish junk placement, we would need the recipient’s observed folder or filtering evidence.
The current DMARC specification (opens in a new tab) explains alignment between an authenticated domain and the visible From domain. An SPF pass and a DKIM pass can both exist without either being aligned, so read the domains as well as the result words.
4. Compare a mailbox message with an application message
Use authorised test recipients and non-confidential content, or suitable existing examples. Keep the recipient environment and relevant message conditions comparable. Record differences rather than treating one successful message as proof that every route works.
| Evidence | Manual mailbox message | Application-generated message |
|---|---|---|
| Visible From | accounts@sender.example | accounts@sender.example |
| Trusted receiver’s reported sending host | outbound.mailbox.example | outbound.mailer.example |
| SPF domain reported by the receiver | sender.example | mailer.example |
| DKIM domain with a reported pass | sender.example | sender.example |
| Record to match next | Mailbox trace using recipient, time and Message-ID | Application sending event plus any provider trace |
| What this comparison establishes | A different reported infrastructure route | An application route to investigate, not a proven cause of spam |
The same From address can therefore sit above different sending infrastructure. Conversely, an application connected to Microsoft 365 may use the same infrastructure as a staff email. In that case, headers may not distinguish the two, and the application event, connected account and sending time become more important.
For finance teams, the invoice email guide explains platform-owned and connected-mailbox sending. If a route changed recently, use the post-migration checks to compare the old and current configuration.
5. Avoid conclusions the headers cannot support
| Observation | Reasonable next step | Unsupported conclusion |
|---|---|---|
| A pass appears in Authentication-Results | Identify the trusted evaluator and the domain tested | The message is genuine, safe or in the inbox |
| A sending IP belongs to a large email provider | Match the route to tenant and application records | That provider caused the spam decision |
| The earliest Received line names a host | Establish whether that part of the trace is trusted | The lowest line always reveals the original sender |
| SPF fails after forwarding | Review the forwarding route, DKIM and receiver handling | The original sender’s SPF must be wrong |
| The domain’s current MX records name Microsoft or Google | Identify the actual outbound route in the received message | Every email from the business uses that provider |
| No message arrived and no received headers are available | Use application errors, bounce details and sending traces | A header analysis can establish the missing message’s placement |
6. Turn the reading into a useful technical handover
Keep the affected message, its exact recipient and time zone, the observed outcome and any matching application event together. State which system supplied each piece of evidence and what remains unknown. If several administrators are involved, name the owner of the next check.
For example: “The recipient gateway records this sending host, and the CRM log contains the matching Message-ID. The message was found in junk, although the gateway reports aligned authentication. The next check is the receiving system’s filtering verdict.” That is a scoped investigation, not a conclusion that authentication or the CRM caused the problem.
For receiving-side interpretation, see the Microsoft 365 guide or Google Workspace guide. A shared sending route may need both your IT service provider and the business application owner. Our accountancy page covers the wider practice-email workflow.
Neotiqa’s email deliverability assessment brings the route, authentication and message evidence into a prioritised technical handover. The assessment is read-only, with implementation or monitoring agreed separately. An initial enquiry needs a description of the problem and systems involved, not confidential headers or mailbox access.
Questions about email header analysis
Can email headers prove which CRM sent a message?
Not always. They can show the reported transport and authentication identities, but several applications may share a provider or use the same connected mailbox. Match the received message to an application event before naming the originating CRM.
Is Return-Path the same as Reply-To?
No. Return-Path records the envelope return address used for delivery-status handling, while Reply-To tells a mail client where normal replies should go. Neither should be substituted for the visible From identity when interpreting DMARC alignment.
Why can SPF pass while DMARC fails?
SPF may pass for a domain that does not align with the visible From domain. DMARC needs an aligned SPF or DKIM pass. Read the tested domains and the trusted receiver’s results together, rather than checking for the word pass alone.
Can I use the headers from a forwarded message?
A normal inline forward has new outer headers. Ask for the original message through your agreed secure process, or for the original received headers. If an original message is attached, distinguish its inner headers from those of the outer forwarding message.
Should I paste business headers into a public analyser or AI tool?
Check your organisation’s data-handling rules first. Headers can expose addresses, internal routes and identifiers, while a full message export may include the body and attachments. Use an approved process and retain an untouched original securely for the investigation.
Several systems sending email, but no clear explanation?
Tell us which messages are affected and which platforms are involved. We can scope an assessment with the people who manage those systems.