Why DMARC reports show sources you don’t recognise
Updated
An aggregate report lists every IP address that sent mail with your domain in the From: header to that receiver, not only the servers you set up. Almost every unfamiliar row is one of four things: a service of yours nobody told you about, forwarding, a mailing list, or someone spoofing your domain. The combination of SPF and DKIM results on the row usually tells you which.
This guide is about sorting the rows. For what each field in the XML means, see How to read DMARC aggregate reports.
Why the receiver reports them
The receiver reports the IP that connected to it. When a message passes through another server first (a forwarder, a list server), that server is what the receiver sees, so the row shows its IP and not yours. And anyone can put your domain in From:, so spoofed mail is reported to you too. That is the point of DMARC reports: they show all mail that claims to be from you.
The four patterns
Read three things per row: the DKIM signing domain (d=) and its result in auth_results, the SPF domain (the envelope_from) and its result, and the volume over several days.
| What the row shows | Most likely | What to do |
|---|---|---|
DKIM pass with d= your domain, SPF fail, an IP you do not know | Forwarding | Nothing. DMARC passes on DKIM. |
DKIM pass with d= a service's domain, SPF pass for that service's bounce domain, DMARC fail, steady volume | Your own service, not set up for your domain | Find the service and set up DKIM with your domain there. |
DKIM fail with d= your domain, SPF fail, small counts from the same few IPs | A mailing list that changed the message | Usually nothing. See Mailing lists. |
No DKIM signature, or one from an unrelated domain, SPF fail or none, IPs spread over many networks, in bursts | Spoofing | Nothing on that source. A quarantine or reject policy is what stops it. |
A fifth case is less common but easy to miss: SPF pass and aligned, no DKIM, from an IP that is yours but not your mail server. That is a web server, a printer or an old application sending directly, allowed by an a, mx or ip4: in your SPF record. It works until the IP changes or the record is cleaned up.
Your own forgotten services
This is the case that matters, because these are real messages that a stricter policy would block. The clues are in the row itself:
- The DKIM
d=domain usually names the service or its sending platform, even when DMARC fails. - The
envelope_fromdomain is often the service's bounce domain, for examplebounces.example.org. - The reverse DNS of the IP often names the platform too. Lookproof's reverse DNS check and IP address info show the PTR name and the network the IP belongs to.
- The pattern: every day, or on the same day each month (invoices, statements), from a small, stable set of IPs.
Then ask the teams that buy software (sales, support, finance, HR). When you find the service, follow the checklist for a new sending service.
Forwarding
A recipient forwards your mail to another address: an alias, an old university account, a rule that sends everything to Gmail. The forwarder sends it on from its own IP, which is not in your SPF record, so SPF fails. If the forwarder does not change the message, your DKIM signature still verifies and DMARC passes. RFC 7960 explains it the same way: for forwarded mail, SPF cannot help and DKIM has to be relied on.
How to recognise it: DKIM passes for your own domain, the IPs belong to well-known mail providers or universities, and the volume follows your own mail (more when you send more). Some receivers mark these rows with the reason forwarded, most do not. As long as you sign with DKIM, forwarding needs no action.
Mailing lists
A list server receives a message from someone at your domain and sends it to every member. Lists that add a [list-name] tag to the subject or a footer to the body break the DKIM signature, and the list server's IP fails SPF. The row then fails both, just like spoofing (RFC 7960, section 3.2.3).
How to tell it from spoofing: the same few IPs, small counts, a reverse DNS name that points to a list host, and colleagues who post to that list. Some list software avoids the problem by rewriting From: to the list's own address when the author's domain has a strict policy; then your domain does not appear in the report for that mail at all.
Some intermediaries add ARC headers (RFC 8617, an experimental standard), which record the SPF, DKIM and DMARC results they saw before changing the message. A receiver that trusts the intermediary can use them to deliver the message despite the DMARC failure, and may report it with the reason local_policy or mailing_list. You cannot control this from your side.
Spoofing
Mail that uses your domain without your permission: phishing, spam, or a backscatter campaign. It looks like this:
- no DKIM signature, or a signature from a domain that has nothing to do with you;
- SPF
failorsoftfailfor your domain, orpassfor an unrelated domain; - IPs spread over many networks and countries, often residential or cheap hosting;
- bursts: thousands of messages over a few days, then nothing.
Do not try to block the IPs or contact the networks. The fix is the policy: with p=quarantine or p=reject, receivers that follow DMARC put these messages in spam or refuse them. The rows keep appearing in your reports, now with disposition quarantine or reject, which is how you see that the policy works. Lookproof's p=none finding and unprotected subdomains finding explain what is still open.
When you still cannot tell
- Look at the same IP across several days and several receivers. Your own services and forwarders are steady; spoofing comes and goes.
- Check whether the volume is large enough to matter. A row of 2 messages a week from an unknown IP is not worth a day of research.
- Keep a short list of the sources you have identified, with the service name and who owns it. Next month's report is then quick to read.
- When every source with real volume is identified and passes, move the policy up: see Moving DMARC from p=none to p=reject.
Senderward checks your domain's DMARC, SPF, DKIM, MX, MTA-STS and TLS-RPT every day and emails you, with the fix, when something gets worse. $7 per domain per month, 14 days free. See pricing