Mail started failing after we published DMARC
If mail started failing after you published DMARC, check the policy you published before changing anything else. A discovered p=none policy must not modify existing mail handling. At p=quarantine or p=reject, receivers are asked to act on messages that have neither an aligned SPF pass nor an aligned DKIM pass. Find the affected sending source and fix that path, rather than assuming the whole domain needs weaker protection.
Publishing a record and a delivery problem happening together don't prove one caused the other. Save a bounce, a message header if you have one, and the DNS values before and after the change. Those give you something specific to investigate.
First, separate monitoring from a handling request
p=none offers no expression of preference. It doesn't turn off the receiver's spam, reputation, or other authentication checks. If you published p=none and mail broke, review other changes made in the same window:
- Did a vendor setup create a second SPF record? See Two SPF records.
- Did another include take SPF over its lookup budget? See SPF permerror: too many DNS lookups.
- Was a DKIM selector deleted, renamed, or published at the wrong name? See DKIM signature not found.
- Did an SPF edit introduce a syntax error or remove a sender?
With quarantine or reject, an existing authentication or alignment problem may now have a delivery consequence. The policies are requests, though: the receiver makes the final handling decision. Quarantine isn't a universal instruction to put mail in a particular folder (What happens to failing mail under each policy).
Read the whole rejection, not just the number
Gmail gives three different explanations for 550 5.7.26:
| Text in the reply | What to investigate |
|---|---|
| "Unauthenticated email from domain-name is not accepted due to domain's DMARC policy." | Authentication and alignment against the From domain, and the policy being applied |
| "This email has been blocked because the sender is unauthenticated. Gmail requires all senders to authenticate with either SPF or DKIM." | Why neither SPF nor DKIM passed; the number alone doesn't identify a DMARC-policy rejection |
| "The (E)MAIL FROM domain... has an SPF record with a hard fail policy (-all) but it fails to pass SPF checks" | SPF for the envelope sender; this is an SPF hard-fail rejection, not proof your DMARC policy caused it |
Google's list also has a 4.7.26 reply, paired with 451, that uses the DMARC-policy wording but adds that "temporary DNS failures prevent authentication." That one points at DNS reachability for your SPF or DKIM records, not a missing sender. The same list also has a 421 4.7.26 reply that uses the unauthenticated-sender wording and says the message has been rate limited, so match the full text, not only the enhanced code.
Google also lists an alignment message under 5.7.32: the From header "isn't aligned with either the authenticated SPF or DKIM organizational domain." Its table pairs that enhanced code with 421. When you quote it in a ticket, use the enhanced code and the text you actually received.
An Outlook.com 550 5.7.515 points to its high-volume sender authentication requirements, not necessarily a rejection requested by your own DMARC policy. Follow Microsoft and Outlook authentication rules for that case.
For Microsoft 365 receiving mail, when Honor DMARC record policy is on (the recommended setting), failing quarantine-policy mail goes to Junk Email, configurable to quarantine, and failing reject-policy mail is rejected during SMTP. Settings and overrides matter. Don't apply that description to every receiver.
Find which source lost its aligned pass
In a message's Authentication-Results, compare the visible From domain (header.from) with:
smtp.mailfrom, the domain SPF checked, and whether SPF passed;header.d, the DKIM signing domain, and whether that signature verified.
One passing, aligned identifier is enough for DMARC. A raw pass for an unrelated vendor domain isn't enough. Use Identifier alignment for SPF and DKIM for the comparison rules, then Third-party mail fails alignment for a third-party sender that needs different authentication settings.
Next, use aggregate reports to group the affected traffic by source. Match those rows to your sender inventory instead of treating every unfamiliar IP as legitimate (A field-by-field walk-through of a DMARC aggregate report, How to inventory sending sources, Reports show sources we do not recognize). Check the particular stream people reported: invoices, newsletters, a website form, or staff mail can take different paths.
If only indirect deliveries are affected, investigate forwarding or mailing-list changes (Forwarded mail fails DMARC). If the From address uses a subdomain, check which policy applies to it (Subdomain policy is not doing what we expected). A parent-policy change can affect more than the address you tested.
Reports aren't a complete incident log. Receivers don't all report, and a message rejected on SPF before the message content is transmitted never reaches DMARC evaluation and never appears in aggregate DMARC reports. Keep the bounce evidence too.
Restore legitimate mail without hiding the problem
Start with aligned DKIM at the affected source. A properly authenticated, aligned Return-Path can also supply the SPF route to a DMARC pass, but it doesn't replace DKIM for domains at p=reject: those domains must not rely solely on SPF and must apply valid DKIM signatures. Reject remains a careful per-domain decision, especially where staff use mailing lists (When you are ready for p=reject).
If critical legitimate mail is being lost while you investigate, a temporary policy step-back may be necessary. Agree on who owns the repair and when you'll reassess it. Reducing a policy affects protection beyond the broken sender, doesn't fix its authentication, and isn't instant because DNS answers are cached. The main-site rollout guide covers moving forward safely after the repair.
Don't add unknown IPs to SPF just to make failures disappear. Don't assume the t=y testing flag is an emergency bypass: receiver support isn't documented in the sources behind this guide.
After the change, send again through the affected service and delivery path. Check the received authentication results and follow subsequent reports. A successful test from your normal mailbox doesn't prove the invoicing platform or a forwarded copy is repaired.
Mistakes worth avoiding
- Publishing DMARC and editing SPF or DKIM in the same change window, then blaming whichever change is easiest to see.
- Moving to quarantine or reject before reading reports for the streams you know about (A safe rollout: p=none to quarantine to reject).
- Adding the vendor's include when its Return-Path isn't your domain. It spends lookups without giving an aligned SPF pass (Third-party mail fails alignment).
- Treating every failing source as yours. Some of them are forgeries (Reports show sources we do not recognize).
- Checking the organizational domain and forgetting subdomains (Subdomain policy is not doing what we expected).
Free DMARC account
Not sure which sender is failing?
Connect your domain's DMARC reports and see every service that sends as your domain, and whether each one passes SPF, DKIM and DMARC.
- Every sending source in one dashboard
- SPF, DKIM and alignment results per sender
- No XML files to read
FAQ
We published p=none and mail started bouncing. Is DMARC the cause?
Unlikely. p=none asks for no change in handling. Look at other DNS changes from the same day, especially SPF and DKIM.
Gmail says 550 5.7.26. What now?
Match the full text against Google's three wordings. Only one of them is about your domain's DMARC policy.
Should we go back to p=none?
Only if legitimate mail is being lost and you can't fix the source quickly. Treat it as temporary, fix the sender, then move forward again.
Will the failed messages show up in our DMARC reports?
Messages that reached DMARC evaluation at a reporting receiver can, with the disposition applied. Messages rejected on SPF before the content was sent won't.