How Do You Move DMARC From p=none to p=reject Without Breaking Email?
Published 2026-11-14 · Reviewed 2026-11-14 · 7 min read · By the Ingenious AI team
Is someone impersonating your domain right now?
Run the free DomainGuard scan: lookalike registrations, MX-armed variants, certificate-transparency and CAA checks for your domain, in one report.
Run my free DomainGuard scanThe short answer
You move DMARC to enforcement in three staged steps: publish p=none and read aggregate reports for at least two weeks, fix every legitimate sender until it passes SPF or DKIM with alignment, then step to p=quarantine at a rising percentage, and only then to p=reject. The record itself takes minutes to change. The work is the inventory: marketing platforms, payroll, your help desk, and the forgotten invoicing tool all send as your domain, and enforcement blocks any of them that still fail alignment. Teams that rush this break real mail. Teams that follow the reports almost never do.
Why p=none protects nobody
A DMARC record at p=none is a monitoring setting, not a defense. It asks receivers to send you reports about mail claiming to be from your domain, but it tells them to deliver everything anyway — including a spoofed invoice sent by a stranger. Attackers know this, and a quick DNS lookup tells them your domain is unenforced before they ever send. That is why a spoofable check flags p=none as an open door: the record existing is not the same as the record protecting you.
The good news is that p=none is exactly where a safe rollout starts. The reports it generates are the map of everything that sends as you. Treat the monitoring phase as an audit you are lucky to get for free, not as a waiting room. Every week you stay at p=none with clean, understood reports is fine. Every month you stay there because nobody opened the reports is exposure you chose.
Step one: inventory every sender in your aggregate reports
DMARC aggregate (rua) reports arrive daily from major receivers as XML files listing every source that sent mail using your domain, with pass and fail counts for SPF and DKIM alignment. Read them for sources, not for totals. You will find your obvious platforms first — your mailbox provider, your CRM, your newsletter tool — and then the surprises: the support desk a colleague connected in 2023, the scanner that emails invoices, the monitoring service that alerts as your domain.
For each source, record three things: what it is, whether it passes SPF alignment, DKIM alignment, or neither, and who owns it internally. Sources passing neither are your fix list. Sources you do not recognise at all are either shadow IT or actual spoofing, and both deserve a direct question this week, not next quarter. Do not advance the policy until every high-volume source is identified and passing at least one aligned check.
Step two: fix alignment, sender by sender
Alignment is where rollouts actually succeed or fail. SPF aligns when the domain in the return-path matches your organisational domain; many third-party senders default to their own return-path, which passes SPF for their domain but not aligned SPF for yours. The fix is usually a custom bounce or return-path subdomain the vendor asks you to delegate or verify. DKIM aligns when the signing domain (the d= value) is your domain, which most reputable senders support once you publish the public key they give you in your DNS.
Work the fix list in volume order. Your biggest sender failing alignment is the one that will hurt at p=reject, so it goes first. After each fix, wait for the next day's reports and confirm that source now shows aligned passes before moving on. Resist changing the DMARC record and a sender configuration on the same day — when something breaks, you want exactly one suspect.
Step three: quarantine at a rising percentage
When reports show your legitimate mail passing, move to p=quarantine with a percentage tag, starting at pct=10. At that setting, receivers send roughly ten percent of failing mail to spam folders and deliver the rest, which gives you a live safety sample: if a real sender was missed, only a fraction of its mail is affected, and your reports and user complaints will name it within a day. Raise the percentage in steps — 10, 25, 50, 100 — spending a few days at each.
Quarantine is also a useful permanent home for subdomains you do not fully control. The sp= tag sets the policy for subdomains separately, so you can enforce p=reject on your main organisational domain while keeping a cautious sp=quarantine on subdomains used by vendors. Being deliberate here is not weakness; it is how large, messy estates enforce without drama.
Step four: reject, then keep watching
p=reject tells receivers to refuse failing mail outright at the SMTP stage. This is the only setting that fully closes exact-domain spoofing: a forged message from your real domain, sent from a server you do not control, simply does not arrive. Publish it once quarantine at 100 percent has run clean for at least a week. Then announce it internally, because the next time someone connects a new tool that sends as your domain without setting up DKIM, its mail will bounce — and that is the system working.
Enforcement is not the end of maintenance. Keep reading aggregate reports monthly, rotate DKIM keys on schedule, and re-run this same staged process for every new domain or subdomain that starts sending. Pair enforcement with lookalike monitoring, because DMARC does nothing about a cousin domain — that gap is covered in our worked example below and in the related guides.
Five actions to take today
- Publish DMARC at p=none with an rua reporting address and collect aggregate reports for at least 14 days.
- Build a sender inventory from the reports and identify an owner for every source sending as your domain.
- Fix SPF return-path or DKIM signing alignment for every legitimate sender, verifying each fix in the next day's reports.
- Step to p=quarantine at pct=10 and raise 10 → 25 → 50 → 100 over at least a week, watching complaints and reports at each step.
- Publish p=reject, announce it internally, and keep a monthly report review so new senders get aligned before they break.
Worked example: Harbor Lane Accounting moves to reject in five weeks
Harbor Lane Accounting, a fictional 22-person firm, started at p=none with 6 domains monitored by DomainGuard. Its first aggregate reports showed 9 sending sources, including a payroll platform nobody had aligned and an old newsletter account still sending monthly. The payroll tool was fixed with a vendor-supplied DKIM key in week two; the newsletter account was retired. The firm stepped to quarantine at 10 percent in week three, caught one misconfigured scanner at 25 percent, and reached p=reject in week five. In the following quarter, receivers reported 1,140 spoofed messages using its exact domain — every one rejected — while its legitimate delivery stayed at 99.2 percent.
FAQ
How long should I stay at p=none before enforcing?
At least two weeks, and longer if your sending is seasonal. You need reports covering every regular sender — payroll cycles, monthly invoicing, and quarterly newsletters included — before you can trust the inventory.
Will p=reject break my email?
It breaks mail from senders that fail both aligned SPF and aligned DKIM. If you fixed those senders during monitoring and quarantine staging, legitimate mail keeps flowing. The staged percentage steps exist precisely to surface stragglers safely.
What is the difference between quarantine and reject?
Quarantine tells receivers to treat failing mail suspiciously, typically by filing it in spam. Reject tells them to refuse it during the SMTP conversation, so it never reaches the mailbox at all. Reject is the stronger and final setting.
Does DMARC stop lookalike-domain phishing?
No. DMARC protects your exact domain only. A lookalike such as a swapped-letter cousin passes its own authentication because the attacker controls its DNS. That is why enforcement pairs with lookalike monitoring.
Can I enforce on subdomains separately?
Yes. The sp= tag sets the policy inherited by subdomains, so you can run p=reject on the organisational domain while subdomains follow a stricter or staged policy of their own.