Skip to content
English
  • There are no suggestions because the search field is empty.

What to do when your email domain is spoofed

A screenshot of a fake invoice isn't enough to go on. If someone says your email domain is being spoofed, get the full headers of the message they received and read two things: the domain in the From address, and the Authentication-Results line. You're trying to find out whether the message used your exact domain (someone@example.com), a lookalike, or only your display name. DMARC can help with the first case. It does nothing for the other two, and changing policy before you've checked can block your own mail.

What spoofing is, and how SPF, DKIM and DMARC fit together, is on our email spoofing page. This article is the checklist for when it's already happening.

Confirm what kind of impersonation you're dealing with

Ask for the original message, not a forwarded rewrite and not a cropped screenshot. Forwarding can change headers. The useful copy is the message source from the mailbox that received it. There's a separate article on how to share those headers safely.

In the headers, look for:

  • From. The domain after @ is what DMARC evaluates. Read that before you trust the display name in front of the address.
  • Authentication-Results. The receiving system writes dmarc=, spf= and dkim= here, and usually header.from= for the domain it checked.
  • Return-Path, or the envelope sender. That's what SPF checked, and it often isn't the visible From address. Envelope From vs Header From covers the difference.

Then put the complaint into one of these cases:

What you see What it usually is What DMARC on example.com can do
Your exact domain in From, and dmarc=fail (no aligned SPF or DKIM pass) Exact-domain spoofing After you enforce, you can ask receivers to quarantine or reject it
A different domain that only looks like yours A lookalike Nothing. That domain has its own policy, or none
Your name in the display name, some other domain in the address (for example Accounts Payable <notices@sender.example>) Display-name spoofing Nothing. DMARC doesn't evaluate display names
Your exact domain, and dmarc=pass Often a compromised mailbox, or a sender you already authorized A policy change won't catch it. Treat this as an account or vendor-access incident

If you can't get headers, you can still check your DNS records and whether reports are arriving. Treat the complaint as unconfirmed until you see a From domain.

Check the SPF, DKIM and DMARC records you already have

Look up the TXT records at example.com and at _dmarc.example.com. You don't need a public checker for this. The DNS records are the source of truth.

You're answering a few practical questions. Is there an SPF record, and is there only one? Two v=spf1 records make SPF fail permanently (Multiple SPF records). For a domain that sends mail, we end the record with ~all, not +all. For a domain that sends nothing, the record should be v=spf1 -all (How to write an SPF record, Softfail vs fail).

Is there a DMARC record at _dmarc.example.com, and does it start with v=DMARC1? If it exists, what's p, and is rua set? p=none without rua asks receivers to change nothing and gives you no reports (Why staying at p=none is a problem).

DKIM keys live at selector._domainkey.example.com. You need the selector, from a real message or from the vendor, before you can check them (DKIM selectors and DNS records).

Publish or fix DMARC so reports start arriving

If there's no DMARC record, publish one at monitoring with a report address:

_dmarc.example.com.  TXT  "v=DMARC1; p=none; rua=mailto:dmarc-reports@example.com"

Replace the mailto with the mailbox or platform address you'll actually read. If reports go to a domain that isn't example.com, that domain has to authorize them (RUA, RUF, and report authorization).

Don't set p=reject on a domain that staff use for everyday mail just because a complaint came in. We'd start at monitoring, every time, unless the domain sends no mail at all. Monitoring is how you see every service that sends as you, including the ones nobody remembered, before you ask receivers to block failures.

If you already have p=none and no rua, add rua first. If reports are already arriving, skip ahead to identifying senders. Staying at p=none forever leaves exact-domain forgeries to each receiver's other filters (Why staying at p=none is a problem).

Reports can take a day or two to start, and not every receiver sends them. If the mailbox stays empty, see Why DMARC reports do not arrive.

Identify your legitimate senders before you tighten anything

Once reports arrive, don't treat every failing row as an attacker. Failing mail usually comes from your own services that aren't aligned yet, from forwarded or list mail, or from unauthorized use of your domain. The way to sort those is in Legitimate mail vs spoofing in reports.

Build the inventory (How to inventory sending sources). Confirm an owner for each recurring source before you change policy. A billing platform, a scanner, or a web form that fails quietly at p=none will fail visibly at quarantine.

The field-by-field reading guide is in DMARC aggregate report walk-through. If a source still looks unfamiliar after that pass, see Reports show sources we do not recognize. Don't add its IP to SPF to make the failure go away. You may be authorizing the spoofer.

Move to p=quarantine, then decide on p=reject

When your legitimate sources pass with alignment, raise p to quarantine on one domain at a time, keep rua, and compare the next reports with the ones you already have. The evidence gates are in A safe rollout from p=none to quarantine to reject. Readiness for each step is in When you are ready for p=quarantine and When you are ready for p=reject.

Don't use pct to stage the change. It's historic. Leave it out of new records, and remove it from old ones (pct is historic).

p=reject is where we'd like most sending domains to end up, but it's a per-domain decision. For a general staff mailbox domain, don't jump there. At reject, every legitimate stream should pass with aligned DKIM, not SPF alone. Parked and non-sending domains are a different case; those can go to reject once you're sure they don't send.

A DMARC policy is still a request. Receivers make the final decision, so reports after enforcement show what those receivers applied, not a guarantee that every path on the internet did the same.

Protect subdomains and domains that shouldn't send

A record on example.com also covers subdomains that don't have their own record. If you omit sp, they inherit p. If you set sp=none while the parent is at quarantine or reject, spoofers can still use names like anything.example.com with no policy behind them. That's a common hole in the middle of an incident (Subdomain policy (sp)).

np covers non-existent subdomains of the organizational domain, a pattern spoofers like. If you don't set np, receivers fall back to sp, then to p.

For a domain that shouldn't send at all (parked, or held for brand protection), this pair is the one we'd publish, once you're sure no system still uses it:

example.com.         TXT  "v=spf1 -all"
_dmarc.example.com.  TXT  "v=DMARC1; p=reject; rua=mailto:dmarc-reports@example.com"

Subdomains don't inherit SPF or DKIM. A subdomain that does send needs its own SPF and DKIM even when it inherits the parent DMARC policy.

What to tell customers and staff

Be precise about what you know. If you only have a screenshot, say you need the full message or the headers before you can confirm it used your domain.

For staff, a short internal note is enough. Don't pay or click through on unexpected requests that appear to come from the company. Check the actual email address, not the display name. Send the original message to whoever handles security. Don't change production DNS on the back of a single complaint.

For customers, tell them how to read the address, and give them a channel they already know: a phone number, a portal, or a mailbox they already use, not a link inside the suspicious message. If the From domain wasn't yours, say so. They may still have recognized your brand. That doesn't mean the mail came from example.com.

Lookalike domains: what DMARC can't do

If header.from is some other domain, your DMARC policy on example.com is never consulted. Any reports for that lookalike, if its owner publishes rua at all, go to them, not to you.

Tightening p on your domain will not stop this. The practical steps are slower and less certain: report the domain to the registrar or host when you can identify them, warn the people who were targeted, and decide whether defensive registrations are worth it for the closest variants. None of that is guaranteed, and this isn't legal advice.

Leave BIMI until after enforcement

A logo in supporting inboxes comes after enforcement, not during the incident. BIMI needs DMARC at quarantine or reject on the organizational domain and its subdomains (sp=none isn't enough), and it doesn't change whether spoofed mail gets through. If you want the visible mark later, follow the right order: SPF, DKIM, DMARC, then BIMI, then What is BIMI and why DMARC comes first and How to move from DMARC enforcement to BIMI.

Common mistakes

  • Flipping p to reject the same day a complaint arrives, before the inventory is done.
  • Adding an unknown IP to SPF because a report showed a failure.
  • Treating a screenshot, or a forwarded rewrite, as proof of the From domain.
  • Assuming a "scam invoice from us" used example.com. It may be a lookalike or a display name.
  • Publishing p=none without rua, then wondering why you can't see the abuse.
  • Raising the parent to quarantine or reject but leaving sp=none.
  • Hunting lookalike domains in your own DMARC reports. They won't be there.
  • Reaching for BIMI as the first fix.
  • Ignoring mail that passes DMARC. A compromised mailbox looks legitimate in reports.

Free DMARC trial

Turn raw DMARC reports into a readable dashboard

Point your rua address to our platform and read the results by sender and source instead of opening XML files.

  • Results grouped by sending source
  • SPF, DKIM and alignment results per sender
  • The original XML reports are still there to download

Start a free DMARC trialTalk to an expert

FAQ

Someone is spoofing my email domain. What should I do first?

Get the full headers and confirm the From domain. Then check whether you already have a DMARC record with rua. Don't change p until you know which case you're in and which senders are yours.

Will publishing p=reject right now stop the spoofing?

Not safely, and not completely. On a domain that sends real mail, a sudden reject can block services you forgot to align. Even at reject, a DMARC policy is a request, and it only applies to your exact From domain. Lookalikes and display-name abuse are unaffected.

Why don't I see the spoofed messages in my DMARC reports?

Possible reasons: you have no rua (or no DMARC record), the receiver doesn't send reports, the mail used a lookalike or only a display name, or it came from a compromised mailbox and passed. Exact-domain failures show up only when a receiver both checks DMARC and sends you aggregate reports.

Can DMARC stop lookalike domains?

No. Your record on example.com is not consulted for any other domain, even one that looks similar.

A customer got a fake invoice that looked like it came from us. Is our mail broken?

Not necessarily. Get the headers. The From domain may be a lookalike, the display name may be the only thing that looks like you, or a real account may be compromised. Start there, not with a DNS change.

Does a failing row in my DMARC report mean I'm being spoofed?

Not by itself. It may be your own misconfigured service or a forwarder. How to sort those sources is in Legitimate mail vs spoofing in reports.

When can we add a BIMI logo?

After DMARC is at enforcement on the organizational domain and its subdomains, and your legitimate mail is passing. It isn't part of the incident response.