Choose a DMARC policy your senders can support
A stricter DMARC policy can affect real mail as well as forged mail. Before changing it, account for every service that sends using your domain.
Compare the handling each policy requests
| Policy | Request for mail that fails DMARC |
|---|---|
| p=none | States no handling preference. It doesn't ask receivers to quarantine or reject mail that fails DMARC. |
| p=quarantine | Treat the message as suspicious. |
| p=reject | Reject the message because its use of the domain is considered unauthorized. |
Treat policy as a request to the receiver
Receivers may make their own decisions about failing mail. Your policy tells them what you'd prefer; it doesn't promise delivery.
These limits are part of RFC 9989.
Tighten the policy after reviewing legitimate mail
- 1
Identify the services that send as your domain
Include the app's sign-up and password messages, support systems and any mailing service. Check each sender's authentication and alignment before changing the policy.
- 2
Review failures while the policy is none
Use receiver reports and your sending service's delivery records to tell forgotten senders apart from unauthorized mail. Fix failures affecting legitimate mail before asking receivers to enforce the policy.
- 3
Move to quarantine when the evidence supports it
Watch for legitimate messages being treated as suspicious. Check the effects of forwarding and mailing lists rather than assuming every failure is forgery.
- 4
Consider reject after resolving the remaining failures
Check delivery results again after the change. Keep the previous policy handy in case legitimate mail starts failing.
Review subdomains before changing the parent policy
An organizational domain's sp tag can set a different policy for existing subdomains. The p policy can apply there too, unless a separate applicable subdomain policy overrides it. Review the subdomains your services use before tightening the parent record.
Should I still use the DMARC pct tag?
Older DMARC guides use
pctto request enforcement for a percentage of failing messages. RFC 9989, published in May 2026, removed that tag after inconsistent receiver behavior.Its replacement,
t=y, asks receivers to apply the policy one level down, sop=rejectis treated as quarantine andp=quarantineas none. Don't treatpctas a dependable way to limit a policy change's effect.
Frequently asked questions
Does p=none mean DMARC is switched off?
No. Receivers can still evaluate DMARC and send the reports you request, so p=none is monitoring, not enforcement.
Does reject stop all impersonation?
DMARC addresses unauthorized use of your exact domain. It doesn't stop someone using a lookalike domain or a misleading display name.
Review the record and delivery symptoms
SPF, DKIM and DMARC
Find out why your emails land in spam, and who else can send mail as you.
DMARC record
Tells receivers what to do with mail that fails, and where to send reports.
Why your site's emails go to spam
Work through authentication, sender setup and the signals a DNS scan can't see.