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

DMARC tags explained: p, sp, np, adkim, aspf, rua, ruf, fo, and t

A DMARC record is a short list of tag=value pairs. Under the current DMARC standard, the tags you'll actually use are v, p, sp, np, adkim, aspf, rua, ruf, fo, and t, plus psd, which most organizations never set. Only v is required. Everything else has a default. Three tags from the older standard, pct, ri, and rf, are now historic. If you find them in an old record, remove them.

Here's a typical record with the tags most domains need:

_dmarc.example.com.  TXT  "v=DMARC1; p=quarantine; sp=quarantine; rua=mailto:dmarc-reports@example.com"

Publishing the record step by step is on our DMARC setup page. How DMARC uses these values is in What is DMARC and how it works. What a complete record looks like, with examples, is in DMARC record explained: syntax, tags and examples.

Current DMARC tags at a glance

Tag Required? Default if missing What it does
v Required, must come first none Must be exactly DMARC1. Otherwise the record is ignored.
p Recommended See below Your policy for failing mail: none, quarantine, or reject.
sp Optional Same as p Policy for subdomains that exist.
np Optional (new) sp if set, otherwise p Policy for subdomains that don't exist.
adkim Optional r DKIM alignment: r relaxed, s strict.
aspf Optional r SPF alignment: r relaxed, s strict.
rua Optional none: no aggregate reports Where to send aggregate reports.
ruf Optional none: no failure reports Where to send failure reports.
fo Optional 0 When to generate failure reports. Ignored without ruf.
t Optional (new) n Test mode. t=y asks receivers to treat failures one level more gently.
psd Optional (new) u Public suffix flag. Leave it out unless you run a public suffix.

The current revision, known as DMARCbis, added np, t, and psd. Receivers must ignore tags they don't recognize, so an unknown or misspelled tag won't break the record, but it won't do anything either. Only v has a fixed position. Putting v and p first is harmless to follow.

v: version

v=DMARC1, first, exactly as written. v=dmarc1 or a record that starts with p= isn't a DMARC record.

p: policy

p states what you'd like receivers to do with mail that fails DMARC:

  • p=none: no preference. Monitoring only; it doesn't change how mail is handled.
  • p=quarantine: treat failing mail as suspicious.
  • p=reject: you consider failing mail clearly unauthorized.

These are requests. Receivers make the final decision, and what quarantine looks like in practice differs between them (What happens to failing mail under each policy). The difference between the two enforcement values is covered in understanding p=quarantine and p=reject.

The current standard lists p as recommended rather than required, a change from the older standard. Always include it anyway. The current standard can be read two ways: one reading treats a record with no p as p=none; the other says that if p is missing (or sp or np is invalid), receivers apply p=none only when a valid rua address is present, and otherwise apply no DMARC processing at all. Don't leave your domain dependent on either reading.

sp and np: subdomain policies

sp sets the policy for subdomains that exist, such as mail.example.com. Leave it out and subdomains get your p value. It only works in the organizational domain's record. In a record published on a subdomain, sp is ignored.

np: non-existent subdomain policy

np covers names that don't exist in DNS at all, such as invoices-2026.example.com when you never created it. Spoofers like those names because there's nothing there to check. Without np, receivers use sp, and without sp, p.

sp=none while p is at enforcement leaves your subdomains unprotected. That's also why BIMI, which needs enforcement on the organizational domain and its subdomains, won't accept it. Inheritance, subdomain records, and np are covered in Subdomain policy (sp).

adkim and aspf: alignment modes

adkim sets how closely the DKIM signing domain must match your From domain. aspf does the same for the SPF (MAIL FROM) domain. r, relaxed, accepts the same organizational domain, so bounces.example.com aligns with example.com. s, strict, needs an identical domain. Both default to relaxed, and they're set independently.

Google's example record includes adkim=s; aspf=s. Don't copy that without checking your senders; relaxed is typically enough. Details are in Relaxed vs strict alignment.

rua and ruf: where reports go

rua lists addresses for aggregate reports, ruf for failure reports. Both take mailto: URIs, comma-separated:

rua=mailto:dmarc-reports@example.com,mailto:dmarc@reports.example.net

Without rua, receivers don't send aggregate reports. Many receivers don't send failure reports at all, whatever you put in ruf. An address on another domain needs that domain to publish an authorization record. All of this is in rua, ruf, and report authorization.

fo: failure reporting options

fo says when you'd like failure reports generated. Receivers may choose to adhere to it, and it's ignored if there's no ruf.

Value Requests a failure report when
0 (default) All mechanisms fail to produce an aligned pass
1 Any mechanism fails to produce an aligned pass
d A DKIM signature fails to verify, aligned or not
s SPF fails, aligned or not

Combine values with colons, such as fo=0:d. 0 and 1 can't be used together. fo=1 without ruf does nothing.

t: testing mode

t=y asks receivers not to apply your published policy, and to apply any special handling they have instead, such as rewriting the From header on mailing lists. You should then expect treatment one level below the specified policy: reject treated as quarantine, quarantine treated as none. It has no effect on reports or on a none policy. t=n, the default, means apply the policy as published.

t=y isn't a percentage dial, and how widely receivers support it since the current standard came out in May 2026 isn't documented. Remove it once you're done testing, and before you rely on your policy for BIMI (BIMI requirements checklist).

psd: public suffix domain flag

psd=y is for operators of public suffixes (registries and similar). psd=n marks a domain as an organizational domain for itself and its subdomains. The default, u, lets receivers work it out. Most organizations never set this tag.

Older tags you may still see

These are from the older standard. The current standard removed them, and they are now historic.

Tag What it did Now
pct Applied the policy to a percentage of failing mail Removed. Remove it from your record.
ri Requested a reporting interval in seconds Removed. Receivers should report at least every 24 hours.
rf Set the failure-report format Removed. The format is defined in the current failure-reporting standard.
!10m-style suffix on URIs Capped report size Obsolete syntax that reporters should ignore.

On pct, operational experience showed that the tag was usually not accurately applied, unless the value specified was either 0 or 100. The one useful case, pct=0, which some intermediaries treated as a signal to rewrite the From header, is what t=y replaces.

You'll still see pct in provider documentation. Google still recommends including it during rollout and shows pct=5 in examples. Microsoft 365 still steps through pct values and describes records in terms of the older standard. Google still requires that a BIMI policy not have a pct value below 100. That advice predates the current standard or still follows the older one. Under the older standard, leaving pct out was the same as pct=100, so removing it doesn't lower your policy.

How to phase in enforcement now that pct is gone is in pct is historic: how to phase in DMARC enforcement now. The staged rollout itself is in A safe rollout: p=none to quarantine to reject.

Mistakes worth avoiding

  • Leaving pct=50 or pct=0 in an old record and assuming your policy is half applied, or not applied. Remove it.
  • Adding fo=1 without ruf. It's ignored.
  • Typos in sp or np values. The current standard can be read two ways: one reading is to ignore a bad tag and use the default; the other is that an invalid sp or np is treated like a missing p (p=none if rua is valid, otherwise no DMARC processing). Implementations already disagree. Check the record after every edit.
  • Copying adkim=s; aspf=s from an example without an inventory of your senders.
  • Leaving out p because the current standard calls it recommended.
  • Leaving t=y in place after testing.

Free DMARC account

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

Create a free DMARC accountTalk to an expert

FAQ

Is pct still valid?

The current standard removed it, and the tag is historic. Some older receivers may still read it. Don't use it for rollout, and remove it from existing records.

What replaced pct?

For the useful pct=0 case, t=y. There's no percentage replacement. Phasing in enforcement is covered in pct is historic: how to phase in DMARC enforcement now.

What's the difference between sp and np?

sp covers subdomains that exist. np covers names that don't exist in DNS.

Do I need fo?

Only if you publish ruf and expect receivers to send failure reports. Many don't.

What does psd mean?

A flag for public-suffix operators. Most organizations never set it.