← Blog
A staged roadmap diagram showing the progression from no DMARC record through p=none, p=quarantine, and p=reject on a dark technical background

· Updated

By DMARCdrift

The DMARC Enforcement Roadmap: From p=none to p=reject

10 min readdmarcpolicyenforcementgetting-started

Moving from p=none to p=reject takes four to eight weeks for most domains: one to two weeks collecting reports, two to four weeks at p=quarantine confirming nothing breaks, then the final move to full enforcement. A DMARC record at p=none does nothing to stop spoofing. It only watches. Real protection requires p=quarantine or p=reject, and each step is gated by one question: are all your legitimate senders aligned?

This is the full roadmap: every stage from no record at all to full enforcement, what each one actually protects you from, and how long the whole thing takes. Each stage links to a deeper guide if you want the detail.

The four stages, at a glance

DMARC enforcement moves through four states, in order:

Stage Policy What it does
0 No record No policy, no reports, no protection
1 p=none Collects reports, enforces nothing
2 p=quarantine Failing mail goes to spam
3 p=reject Failing mail is blocked outright

The mistake almost everyone makes is stopping at stage 1. Our research found that 73.8% of well-known domains have a DMARC record but only 60.4% actually enforce one, the gap is entirely domains stranded at p=none, believing they are protected when they are not.

How to find your current stage

Before you plan the next step, locate where you are. Check your DMARC record with the DMARC record checker and read the p= tag:

  • No record returned: you are at stage 0. Start with a record.
  • p=none: stage 1. You are collecting data but enforcing nothing. The next step depends on whether your senders are aligned.
  • p=quarantine: stage 2. You have real protection, with a safety net. Confirm your alignment is high and stable before the final move.
  • p=reject: stage 3. You are fully enforced. Your job now is to keep it that way.

One detail trips people up: a policy at p=quarantine or p=reject combined with t=y is applied one level softer, so p=reject; t=y is handled as quarantine and p=quarantine; t=y as none. A legacy pct= below 100 has a similar softening effect at older receivers. The tag matters as much as the policy. If your record has t=y or a pct= below 100, you are partway through a stage, not finished with it.

Stage 0: No DMARC record

If you have no DMARC record, receiving servers have no instruction for what to do with mail that fails authentication, and you get no reports about who is sending under your name. Anyone can spoof your domain and you will never know. This is the worst place to be, and it is also the easiest to fix.

The full consequences of having no DMARC record come down to two losses: no protection and no visibility. The fix is a single DNS record. You can generate a starter record in a minute and confirm it with the DMARC record checker.

Start at p=none deliberately. Publishing p=reject on day one, before you know your sending sources, is how legitimate mail gets blocked.

Stage 1: p=none (monitoring mode)

p=none tells receivers to take no action on failures but to send you aggregate reports. This is your data-gathering phase. The reports reveal every source sending mail as your domain: your ESP, your transactional provider, your support desk, that marketing tool someone connected two years ago.

The danger of p=none is comfort. It feels like progress, the reports arrive, nothing breaks, and domains sit there for years. But p=none blocks nothing. A spoofer impersonating your domain sails straight through.

The goal of this stage is to reach the point where every legitimate source is authenticating correctly. The detail of what p=none means and when you're ready to move past it is its own guide, but the short version: stay long enough to identify and fix every sender, and no longer. Most domains should spend one to two weeks here, not one to two years.

Stage 2: p=quarantine (the first real enforcement)

p=quarantine is the first policy that actually does something. Mail that fails DMARC alignment is delivered to the spam folder instead of the inbox. This is the safety net before full enforcement: if you missed a legitimate sending source, the cost is a message in spam, not a deleted message.

Move to p=quarantine once your aggregate reports show all known senders passing alignment consistently. A common threshold is 95% or higher aligned mail sustained for two consecutive weeks. Use the DMARC policy simulator to see what enforcement would do to your current mail stream before you commit.

Before full quarantine, run a short soft test: publish p=quarantine; t=y; pct=0. t=y asks receivers to apply the policy one level softer, so failing mail is still handled as p=none while the enforcing record is live. Both tags are there because older RFC 7489 receivers read pct and ignore t, while RFC 9989 receivers read t and ignore pct. When your reports stay clean, remove the pair to enforce. The old pct=25 ramp no longer works because RFC 9989 removed pct; the RFC 9989 overview covers what else changed.

Quarantine vs reject across the scan

p=quarantine asks receivers to filter failing mail; p=reject asks them to refuse it. Receiver policy still determines the outcome. Both can affect legitimate mail that fails alignment, so investigate senders before advancing. RFC 9989 changed the DMARC policy tags, removing the old pct rollout mechanism; the testing approach above uses t=y; pct=0 for newer and older receivers.

Our November 2026 DMARC adoption scan counted the following fully enforced policies in its curated sample. These are domain counts, not mail volumes.

Segment Quarantine Reject
State government 10 15
Tranco 500 67 228
US federal government 30 1,070
Top universities 11 17
Top banks 1 23
Cybersecurity companies 8 32
Fortune 100 15 79
Hospital systems 3 20
Media companies 4 21
Top SaaS 22 75

Stage 3: p=reject (full enforcement)

p=reject is the destination. Mail that fails DMARC alignment is blocked at the receiving server and never delivered anywhere. This is what actually stops someone from spoofing your domain to your customers.

The decision to move from quarantine to reject is fundamentally a deliverability decision as well as a security decision: you are trading a small risk to your own mail for real protection against impersonation. The practical threshold is 98% or higher sustained alignment for a full month at full p=quarantine (no t=y), with confidence that there are no remaining blind spots. If you want one more safety step, publish p=reject; t=y; pct=0 first, which receivers handle as quarantine, then remove the pair.

Once you reach p=reject, the work shifts from configuration to vigilance. New sending sources, expired DKIM keys, and SPF changes can all reintroduce failures, and at p=reject a failure means blocked mail. This is where continuous monitoring stops being optional.

The failure modes after enforcement are predictable. A team signs up for a new marketing tool and sends through it without aligning the domain. A DKIM key rotates and the new selector never gets published. Someone edits the SPF record and pushes it past the ten-lookup limit, silently breaking SPF for every sender. None of these announce themselves. At p=none they would show up as a line in a report nobody reads; at p=reject they show up as mail that quietly stops arriving. The difference between catching them in a day and discovering them when a customer complains is whether something is watching the reports.

The deadline that forced everyone's hand

For years, DMARC enforcement was something diligent teams did and everyone else ignored. That changed in February 2024, when Google and Yahoo began requiring DMARC for bulk senders, anyone sending more than 5,000 messages a day to their users.

The Gmail and Yahoo sender requirements only mandate a DMARC record with a reporting address, which means p=none technically passes. But the deadline pushed millions of domains onto the roadmap for the first time, and the ones that understood the difference between having a record and enforcing one moved straight through to quarantine and reject.

Why most domains stall, and how not to

The enforcement gap exists because p=none is comfortable and the next step feels risky. Moving to enforcement means accepting that you might affect real mail, and without good reporting that feels like flying blind.

It does not have to. Every move up the roadmap is gated by one question: are all my legitimate senders aligned? If you can answer that with confidence, the next stage is safe. If you cannot, you are not ready, and the answer is in your aggregate reports, if someone is reading them.

Putting a calendar on it

The roadmap is shorter than people fear. A domain with a handful of known senders can run it in four to eight weeks:

  • Week 0: publish p=none with a reporting address. Reports begin arriving within 24 to 48 hours.
  • Weeks 1–2: identify every sending source in the reports and fix any that fail alignment. This is the only stage whose length depends on you.
  • Weeks 2–4: run a short soft test at p=quarantine; t=y; pct=0, then move to full p=quarantine and confirm nothing legitimate lands in spam.
  • Week 4–8: once alignment holds above 98% for a full month, move to p=reject, optionally via p=reject; t=y; pct=0 first.

Complex domains, with many ESPs, forwarders, acquired brands, and legacy senders, take longer, sometimes several months. But the calendar is driven entirely by how long it takes to reach stable alignment. The policy changes themselves are instant; the readiness is what you are waiting on.

That is the entire job of monitoring: turn the raw XML into a clear, current answer to "is anything failing, and is it mine?" so each step up the roadmap is a decision backed by data instead of a leap of faith. DMARCdrift ingests your reports automatically, tracks alignment across every sending source, and alerts you when something drifts, so you reach p=reject and stay there.

The roadmap is not long. For most domains it is a few weeks of attention, not a project. The hard part was never the records. It was knowing when to take the next step.

Updated September 28, 2026: replaced pct= ramping with the RFC 9989 (DMARCbis) t=y; pct=0 soft test in the rollout steps.

FAQ

How long does it take to get to p=reject?

For a domain with a handful of known sending sources, four to eight weeks is realistic: one to two weeks collecting reports at p=none, two to four weeks at p=quarantine confirming nothing breaks, then the move to p=reject. Complex domains with many ESPs, forwarders, and legacy senders can take several months. The timeline is driven by how long it takes to reach consistent alignment, not by a calendar.

Do I have to go through p=quarantine, or can I jump straight to p=reject?

You can technically jump from p=none to p=reject, but you shouldn't. p=quarantine is a safety net: failing mail lands in spam instead of being deleted, so a missed sending source costs you a spam folder rather than lost mail. Spend at least a few weeks at p=quarantine to catch blind spots before full enforcement.

Is p=none good enough to satisfy Gmail and Yahoo?

Yes, technically. Gmail and Yahoo's bulk sender rules require a DMARC record with any policy, including p=none, as long as it has a reporting address. But p=none provides zero protection against spoofing. It satisfies the requirement without delivering the security benefit, which is why most guidance treats p=none as a starting line, not a destination.