Introduction
A DMARC record example is only useful if it matches the client's situation. Below are copy-ready DMARC records, all valid under RFC 9989, for the setups MSPs meet most: new monitoring, Microsoft 365, Google Workspace with an ESP, enforcement, subdomains, parked domains and external reporting, each with the matching SPF line.
Every record here is published as a TXT record at _dmarc (for example _dmarc.example.com). If you need a tag-by-tag explanation first, read what a DMARC record is and how to read it; for the DNS steps, see how to create a DMARC record step by step. The wider context lives in the DMARC guide.
Key takeaways
- RFC 9989 defines these tags:
v,p,sp,np,adkim,aspf,rua,ruf,fo,tandpsd.pct,riandrfare removed, so none of the examples use them. v=DMARC1must be the first tag, and a domain may publish only one DMARC record at_dmarc.- Roll out by policy level:
p=none, thenp=quarantine, thenp=reject, witht=yas an optional testing step. - Parked domains get
p=reject, SPF-alland a null MX on day one. - A
ruaaddress on another domain only works if that domain authorises it with a_report._dmarcrecord.
Which tags can a DMARC record use under RFC 9989?
A DMARC record can only use the tags registered by RFC 9989: v, p, sp, np, adkim, aspf, rua, ruf, fo, t and psd. Receivers ignore unknown tags. The old pct, ri and rf tags are listed as historic and are no longer processed.
| Tag | Values | Default |
|---|---|---|
v |
DMARC1 (must be first) |
Required |
p |
none, quarantine, reject |
Recommended |
sp |
Policy for existing subdomains | Falls back to p |
np |
Policy for non-existent subdomains | Falls back to sp, then p |
adkim / aspf |
r (relaxed) or s (strict) |
r |
rua / ruf |
Comma-separated URIs, usually mailto: |
No reports |
fo |
0, 1, d, s |
0 |
t |
y or n |
n |
psd |
y, n, u |
u (only public suffix operators need y) |
The examples below use example.com for the client and example.net for a report platform. The generator guide covers the syntax rules behind them.
What DMARC record should a new domain start with?
A new or unaudited domain should start at p=none with an aggregate report address. This changes nothing about delivery but starts the flow of aggregate reports, so you can find every legitimate sender before enforcing.
_dmarc.example.com TXT "v=DMARC1; p=none; rua=mailto:dmarc@example.com"
Use this on day one of any client onboarding; a p=none record without rua tells you nothing. See what p=none means for how long to stay here.
What does a DMARC record look like for a Microsoft 365-only domain?
A domain that sends only through Microsoft 365 needs Microsoft's SPF include and a DMARC record that starts at monitoring. Microsoft recommends -all for Microsoft 365 domains because DKIM and DMARC are also expected, according to Microsoft Learn.
example.com TXT "v=spf1 include:spf.protection.outlook.com -all"
_dmarc.example.com TXT "v=DMARC1; p=none; rua=mailto:dmarc@example.com"
Use this when the client has no other sending services; once DKIM is enabled for the custom domain and reports are clean, move up the policy ladder. The Microsoft 365 implementation guide covers the DKIM CNAMEs.
What does a DMARC record look like for Google Workspace plus an ESP?
For Google Workspace plus a marketing platform, list both senders in one SPF record and keep a single DMARC record. Google's own example combines _spf.google.com with Mailchimp's include, as shown in Google's SPF setup help.
example.com TXT "v=spf1 include:_spf.google.com include:servers.mcsv.net ~all"
_dmarc.example.com TXT "v=DMARC1; p=none; rua=mailto:dmarc@example.com"
Use this when the ESP signs with DKIM for example.com; many ESPs use their own envelope domain, so SPF alone often will not align and DKIM is what passes DMARC. The Google Workspace guide has the full checklist.
What do quarantine and reject records look like?
Enforcement records only change the p value. Move from p=none to p=quarantine once reports show every legitimate source passing aligned SPF or DKIM, then to p=reject after a stable period with no legitimate failures.
v=DMARC1; p=quarantine; rua=mailto:dmarc@example.com
v=DMARC1; p=reject; rua=mailto:dmarc@example.com
Use quarantine as the first enforcement step and reject as the end state. There is no percentage ramp: RFC 9989 removed pct, so each level applies to all failing mail. Moving from monitoring to enforcement safely describes the evidence to collect between steps.
How do you use the t=y testing flag?
The t=y flag asks receivers to apply the policy one level below the published one. p=quarantine; t=y is treated as none, and p=reject; t=y as quarantine. It does not change reporting and has no effect at p=none.
v=DMARC1; p=reject; t=y; rua=mailto:dmarc@example.com
Use this when you want receivers to see the target policy while failing mail still only receives the softer treatment; remove t=y once reports confirm it is safe. New DMARC tags explained covers the edge cases.
How do you set a different policy for subdomains?
Add sp= when existing subdomains need a different policy from the organisational domain. A common pattern keeps the main domain at monitoring while locking down subdomains that should never send, or the reverse during a staged project.
v=DMARC1; p=quarantine; sp=reject; rua=mailto:dmarc@example.com
Use this when the apex still has senders to fix but no subdomain sends mail; any subdomain that does send can publish its own _dmarc record, which overrides sp. More patterns are in DMARC for subdomains.
How do you block mail from non-existent subdomains?
Use np=reject to reject mail from subdomains that do not exist in DNS, such as invoices.example.com invented by an attacker. RFC 9989 says np applies only to non-existent subdomains, not to existing ones or the domain itself.
v=DMARC1; p=none; np=reject; rua=mailto:dmarc@example.com
Use this even while the apex is still at p=none, because no legitimate mail can come from a name that does not exist. See non-existent subdomains and the np tag.
What records should a parked domain publish?
A parked or non-sending domain should publish p=reject, an SPF record that authorises nothing and a null MX. A null MX is an MX record with preference 0 and the exchange ., defined in RFC 7505.
example.org MX 0 .
example.org TXT "v=spf1 -all"
_dmarc.example.org TXT "v=DMARC1; p=reject; sp=reject; rua=mailto:dmarc@example.com"
Use this for every defensive registration in a client portfolio; there is nothing to break, so no monitoring phase is needed. Should parked domains have a DMARC record? explains the wider checklist.
What record suits a subdomain used only for bulk mail?
A bulk mail subdomain gets its own DMARC record at _dmarc under that subdomain, which overrides the parent's sp. This lets marketing mail follow its own rollout without touching the policy for staff mail on the main domain.
_dmarc.news.example.com TXT "v=DMARC1; p=none; rua=mailto:dmarc@example.com"
Use this when the client sends newsletters from news.example.com; add SPF at the subdomain only if the ESP uses it as the envelope sender. Subdomain or separate domain for bulk email? helps decide the structure.
How do you send reports to an MSP or report platform?
Point rua at the platform's address. When that address is on a different domain, the receiving domain must publish an authorisation record, or receivers will not send the reports. RFC 9990 defines the name as the policy domain, then _report._dmarc, then the report domain.
_dmarc.example.com TXT "v=DMARC1; p=none; rua=mailto:reports@example.net"
example.com._report._dmarc.example.net TXT "v=DMARC1"
Use this whenever the reports go to an MSP-run or third-party address; the platform normally publishes the second record, often as a wildcard. See authorising external DMARC report destinations.
Can a DMARC record send reports to more than one address?
Yes. Separate rua URIs with a comma, each with its own mailto: prefix. RFC 9990 says receivers must attempt delivery to every listed URI, so each address receives a full copy.
v=DMARC1; p=quarantine; rua=mailto:dmarc@example.com,mailto:reports@example.net
Use this when the client wants an internal copy alongside the MSP's platform, for example during a handover. Each external address still needs its own authorisation record.
What does the UK government example look like?
GOV.UK's government email guidance starts public sector domains at p=none with reports sent to the NCSC's reporting address plus an internal mailbox. Its record for domains that do not send email goes straight to p=reject.
v=DMARC1;p=none;rua=mailto:dmarc-rua@dmarc.service.gov.uk,mailto:dmarc@example.gov.uk
The NCSC's reject spoof emails page still shows an example with a pct value of 50 and a percentage ramp. RFC 9989 removed pct, so leave it out and use p=quarantine, then p=reject (with t=y if you want a test step). The NCSC address is for UK public sector bodies; what the NCSC recommends has the background. To publish one of these records at a UK registrar, adding a DMARC record at 123-reg walks through the Name and Value fields.
Which DMARC records look right but are wrong?
These mistakes pass a quick glance but cause receivers to ignore the record or part of it. RFC 9989 discards every record when more than one is returned, and ignores a record whose first tag is not v=DMARC1.
| Looks right | What goes wrong | Fix |
|---|---|---|
Two TXT records at _dmarc.example.com |
All DMARC records are discarded | Merge into one record |
p=none; v=DMARC1; rua=... |
v is not first, record ignored |
Put v=DMARC1 first |
v=DMARC1; p=reject at example.com |
Receivers only look at _dmarc |
Publish at _dmarc.example.com |
p=quarantine plus a leftover pct tag set to 25 |
pct ignored, policy applies to all failing mail |
Remove pct |
rua=dmarc@example.com |
Not a URI, no reports sent | Use rua=mailto:dmarc@example.com |
“v=DMARC1; p=none” pasted from a document |
Smart quotes become part of the value | Retype with straight quotes or none |
Common DMARC errors covers more failure modes.
How should an MSP manage these records across many clients?
Treat the examples as templates per client type: Microsoft 365, Google Workspace, parked, bulk subdomain. Record which template each domain uses and its current policy level, then change records through a documented change-management policy.
Standardising on one report address per platform keeps authorisation simple. DMARC AI processes aggregate reports per client domain and its AI assistant can check a domain's live DNS records, which helps confirm a published record matches the intended template.
Frequently asked questions
Do I need to include adkim and aspf in every record?
No. Both default to relaxed (r), which suits most clients, because relaxed alignment accepts subdomains of the same organisational domain. Only add adkim=s or aspf=s when you know every sender signs or sends with the exact From domain, otherwise legitimate mail from subdomains will fail.
Should I add ruf and fo to these examples?
Usually not. Many large receivers do not send failure reports, and those that do may include message content, which raises privacy questions. Add ruf and fo=1 only if you have a process to handle the reports securely. Aggregate reports via rua cover routine monitoring.
Does a record still work if it contains pct?
The record still works, but receivers implementing RFC 9989 ignore pct. A p=quarantine record with pct set to 10 is treated as full quarantine on all failing mail, which may surprise a client expecting a 10 percent trial. Remove the tag and use t=y if you want a softer test.
Do quotes matter when I paste a DMARC record into DNS?
Most DNS panels add straight quotes for TXT values automatically, so adding your own is optional. Curly smart quotes from word processors are a problem because they become part of the value and break the v=DMARC1 match. Paste from a plain text editor and check the result with dig.