schedule 12-min read

“DMARC Quarantine/Reject Policy Not Enabled”: What the Warning Means

DMARC AI editorial team · Last updated

Why scanners say “DMARC quarantine/reject policy not enabled”, which record states trigger it, and how to clear the warning safely without blocking mail.

01

Introduction

"DMARC quarantine/reject policy not enabled" means a scanner looked up your domain's DMARC record and found no enforcement: the policy is p=none, the record is missing or broken, or a subdomain policy or test flag weakens it. The domain can still be spoofed. The fix is a staged move to p=quarantine, then p=reject, driven by aggregate reports.

The warning is usually accurate, but it is not an emergency, and the worst response is to flip the record to p=reject the same afternoon. This page explains what each kind of scanner is checking, which record states trigger the finding, how to answer it for a client or an insurer, and how an MSP can triage a whole portfolio. For the policies themselves, see DMARC policy explained; for the protocol as a whole, the DMARC guide.

02

Key takeaways

  • The warning means receivers have no instruction from you to quarantine or reject mail that fails DMARC, so exact-domain spoofing still works.
  • Common triggers: p=none, no record, an invalid or duplicated record, sp=none on subdomains, and t=y, which RFC 9989 treats as one level below the published policy.
  • Google and Yahoo accept p=none for bulk senders, so the warning is about spoofing protection, not deliverability.
  • The NCSC states that a reject policy on all your domains is the best way to prevent spoofing.
  • Clear it in stages (none, then quarantine, then reject), with exit criteria from aggregate reports. There is no pct ramp under RFC 9989.
03

Which scanners raise the "policy not enabled" warning?

Four kinds of tool raise it, and they all read the same public DNS record: _dmarc.yourdomain as a TXT lookup. None of them can see your mail flows. They judge the published policy only, which is why the finding looks identical whether a domain sends ten messages a day or ten million.

Scanner type Where you see it What it usually checks How it scores the finding
Security ratings and vendor-risk platforms A supplier or client's risk report about your organisation Apex domain and sometimes subdomains found in public DNS Often a deduction or "medium" issue until p=quarantine or p=reject
External attack-surface and penetration test scans Pen test reports, managed vulnerability scans Every domain in scope, including parked domains Listed as a misconfiguration with remediation advice
DNS and email health checkers Free web checkers, monitoring dashboards One domain at a time, record syntax and tags Pass, warning or fail per tag
Cyber insurance questionnaires and scans Renewal forms, broker portals, insurer pre-bind scans Usually the main email domain A yes/no question about whether DMARC is enforced

The wording varies ("DMARC policy not enabled", "DMARC not enforced"), but the meaning is the same. Free DMARC checkers: what they show and what they miss explains the limits of record-only checks in more detail.

04

What exactly triggers the warning?

Any record state where receivers would not quarantine or reject failing mail. The obvious one is p=none, but several less visible states produce the same finding, and some scanners check them while others do not. Confirm the live record with our DNS lookup tool before deciding which applies.

Record state What a receiver does under RFC 9989 Typical scanner result
p=none No policy applied to failing mail; reports only Policy not enabled
No _dmarc record No DMARC processing at all DMARC missing (often a separate, higher-severity finding)
Record with no p= tag Treated as if it said p=none (Section 4.7) Policy not enabled
Invalid p= value, valid rua= Treated as p=none (Section 4.10.1) Policy not enabled or record invalid
Invalid p= value, no rua= No DMARC processing (Section 4.10.1) Record invalid
Two DMARC records at the same name All discarded (Section 4.10) Multiple records or DMARC missing
p=reject; sp=none Apex enforced, existing subdomains not Subdomain policy not enabled
p=quarantine; t=y Receivers apply none May be flagged as not enforced
p=reject; t=y Receivers apply quarantine Usually passes, sometimes flagged as test mode

The source for each behaviour is RFC 9989, the current DMARC standard. Section 4.7 says a record without a p tag "is treated as if it included 'p=none'". Section 4.10 says that if multiple DMARC policy records are returned, "they are all discarded".

If the record is missing entirely, start with how to fix "No DMARC record found" and come back here once p=none is published.

Why does a subdomain policy trigger it when the apex is at reject?

Because sp=none switches enforcement off for every existing subdomain. RFC 9989 says the sp tag applies only to subdomains, and that the p value applies to subdomains when sp is absent. A record such as this one protects clientco.co.uk but not invoices.clientco.co.uk:

_dmarc.clientco.co.uk.  TXT  "v=DMARC1; p=reject; sp=none; rua=mailto:dmarc@clientco.co.uk"

Scanners that enumerate subdomains will flag each one. Remove sp=none once subdomain senders are authenticated. DMARC for subdomains and the np tag for non-existent subdomains cover the subdomain side.

Why can t=y trigger it?

Because t=y is the RFC 9989 test mode flag, and receivers apply one level below what you published. Under Section 4.7, p=quarantine; t=y means failing mail gets none, and p=reject; t=y means it gets quarantine. A scanner that understands the tag may therefore report p=quarantine; t=y as not enforced. That is correct: it is a rehearsal, not protection. New DMARC tags explained: np, psd and t has the full definition.

05

Why does the warning matter if mail is being delivered?

Because delivery and protection are different questions. A domain at p=none can meet mailbox provider rules and still be spoofed. Enforcement is the part of DMARC that tells receivers to stop mail that impersonates your domain, and p=none gives you reports about that mail without stopping any of it.

The large receivers set a floor, not a target. Google's sender guidelines treat a missing DMARC record as a problem for bulk senders but list the minimum as p=none. Yahoo's sender best practices require bulk senders to "publish a valid DMARC policy with at least p=none". So a domain can satisfy both and still carry this warning. Google and Yahoo sender requirements covers those rules.

The UK guidance goes further. The NCSC's reject spoof emails stage says: "Having a DMARC policy of 'reject' on all of your domains is the best way to prevent spoofing of your email." Its recommended implementation plan ends there too. Why DMARC enforcement matters more than monitoring makes the broader case. That page still describes a percentage step, which RFC 9989 has since removed. What does the NCSC recommend for DMARC summarises the rest.

06

How do you clear the warning safely?

Move through the policy levels in order and leave each one only when the aggregate reports meet its exit criteria. RFC 9989 (Section 5.1) describes the same sequence: start at p=none with an rua address, analyse the reports, and fix unaligned legitimate streams before any move to enforcement.

  1. Confirm reporting. The record needs an rua= address that someone, or a platform, actually reads. Without it, receivers send no aggregate reports and you have no evidence.
  2. Inventory senders at p=none. Name every source in the reports. DMARC client discovery walks through it.
  3. Exit p=none when every legitimate sender passes with aligned SPF or DKIM, no new senders have appeared for a week, and the data covers at least 14 days and one full business cycle.
  4. Publish p=quarantine, optionally rehearsed first with t=y. Hold for at least 14 days with no legitimate mail failing.
  5. Publish p=reject, optionally rehearsed with t=y, then remove t=y. The warning clears at whichever level the scanner accepts, but the job finishes at reject.

The gates and an example timeline are in how to move from DMARC monitoring to enforcement safely. The record at the end looks like this:

_dmarc.clientco.co.uk.  TXT  "v=DMARC1; p=reject; rua=mailto:dmarc@clientco.co.uk"

There is no percentage setting to stage the move: RFC 9989 removed the pct tag, as explained in pct has been removed.

07

How should you answer the warning for a client or an insurer?

Answer with the current state, the plan and a date, not with a defence of p=none. A reviewer reading a ratings report or an insurance form wants to know whether the gap is known and owned. A short, factual statement usually does more than a hurried DNS change.

For a client, a reply in this shape works:

  • Finding: the DMARC record for the domain is at p=none, which is monitoring mode.
  • Why it is not yet enforced: sending sources are being confirmed from aggregate reports so legitimate mail is not blocked.
  • Plan: move to p=quarantine once all legitimate senders pass with alignment, then p=reject.
  • Target dates and owner: for each step, named in the change ticket.

For an insurance questionnaire, answer the question literally. If it asks whether DMARC is "enforced" or set to quarantine or reject, and the record says p=none, the answer is no. Add the remediation plan in the comments field if there is one. Do not tick yes on the strength of a planned change. If the insurer runs its own external scan, it will see the live record anyway.

If the domain sends no mail at all, the fast answer is correct: publish p=reject straight away. Should parked domains have a DMARC record gives the full record set.

08

How should an MSP triage this warning across a portfolio?

Sort every client domain by its live record and its report data, then work the queue from the quickest safe wins down. The same finding is usually present across many tenants, so fix it as a portfolio exercise.

Bucket How to recognise it Action Typical risk of change
Non-sending or parked domain No mail in reports, no MX or a null MX Publish p=reject now Very low
p=none with clean reports All legitimate senders aligned for 14+ days Move to p=quarantine this change window Low
p=none with failing legitimate senders Known senders failing alignment Fix SPF or DKIM per sender first Medium if rushed
p=none with no rua No reports reaching anyone Add reporting, restart the clock High if enforced blind
Apex enforced, sp=none Subdomain flagged by scanner Authenticate subdomain senders, drop sp=none Low to medium
t=y left in place Record carries t=y after the rehearsal Remove t=y once reports stay clean Low
Missing, duplicate or invalid record Scanner says missing or invalid Publish one valid record at p=none with rua Low

How MSPs should audit client domains for DMARC risk covers the wider audit, and the DMARC change-management policy for MSP portfolios covers approvals and rollback.

09

What should you not do when you see this warning?

Do not jump straight to p=reject without reports. That clears the scanner finding and can bounce invoices, payroll mail and helpdesk tickets from any sender you have not authenticated. RFC 9989 Section 5.1.6 says shortcomings in legitimate mail streams "MUST be addressed prior to any attempt" to publish an enforcement policy.

Other mistakes that come up with this warning:

  • Adding a second DMARC record next to the old one. Receivers discard both.
  • Reaching for the old pct tag. It is gone from the standard and gives no staging.
  • Leaving t=y in place after the rehearsal, so the domain runs one level below its published policy.
  • Fixing only the apex while sp=none leaves subdomains open.
  • Answering an insurer "yes" before the record actually says p=quarantine or p=reject.

Common DMARC errors and how to fix them covers other record faults, and how MSPs can avoid breaking client email with DMARC covers rollout incidents.

10

Frequently asked questions

Is "DMARC policy not enabled" the same as "no DMARC record"?

Not quite. Most scanners use "no DMARC record" when the _dmarc lookup returns nothing, and "policy not enabled" when a record exists but does not enforce. Some tools merge the two. Both mean failing mail is not quarantined or rejected, but a missing record also means you get no aggregate reports.

Does p=quarantine clear the warning, or do I need p=reject?

It depends on the scanner. Many accept p=quarantine as enforcement because receivers treat failing mail as suspicious. Others score p=reject higher or flag quarantine as partial. Either way, quarantine is a step, not the end: the NCSC recommends reject on all domains.

Will moving to p=reject hurt deliverability?

Not for mail that passes DMARC. Enforcement only affects mail that fails, so legitimate senders that pass with alignment are unchanged. The risk is a legitimate sender you missed, which is why the reports gate each move. Google and Yahoo require at least p=none for bulk senders, and enforcement meets that.

How long does it take to clear the warning safely?

For a domain with clean monitoring data, typically several weeks: at least 14 days at quarantine before reject, plus any rehearsal with t=y. Domains with many senders take longer. Parked domains can go to p=reject the same day because they send nothing.

Does the warning apply to subdomains I never use?

It can, if a scanner finds them in DNS. Existing subdomains inherit p unless sp overrides it, and non-existent ones follow np, then sp, then p. Avoid sp=none unless a subdomain genuinely needs it, and consider np=reject.

11

Next steps

Look up the live record for the domain the scanner flagged, find its row in the trigger table above, and check whether you have at least 14 days of aggregate reports. If you do, the exit criteria tell you whether the next policy level is safe. If you do not, add rua= today and start the clock.

For MSPs managing many client domains, DMARC AI processes RFC 9990 aggregate reports per tenant and gives AI-assisted guidance on which senders still fail alignment, which is the evidence each policy move depends on.

Related articles

Related tools

Ready to Implement?

Get authenticated mail moving in minutes: start free, book a guided demo, or talk to the team about your stack.