What breaks the week you set DMARC to reject

Reject is the point of DMARC. It is also when every forgotten sender becomes your problem.

Moving DMARC to reject stops spoofed mail from being delivered as you, and it also stops legitimate systems that send with your domain but fail alignment. Read aggregate reports until the only failing sources are ones you are willing to break, then flip.

Why reject exists

DMARC without reject is mostly visibility. You learn who sends as your domain. Spoofers still land in inboxes when receivers treat p=none as a suggestion.

Reject is the policy that makes the DNS record a control. Receivers that honour it will refuse mail that fails authentication and alignment under your domain. That is the outcome worth the work.

It is also the week your CRM, ticket system, marketing platform, and that one legacy scanner start failing delivery if they were never aligned.

The senders you forgot

Aggregate reports are unkind in a useful way. They show ESP platforms finance never mentioned, a subsidiary’s newsletter tool, a printer that emails scans, a contractor’s outreach sequence still signed with your domain.

None of these feel urgent until reject turns “delivered” into “rejected”. Inventory from memory will miss them. Inventory from two to four weeks of reports will not.

For each source, decide: fix alignment, move it to a subdomain with its own policy, or shut it down. “We’ll deal with it later” is how you roll back reject under pressure.

Alignment, not just authentication

People confuse SPF or DKIM passing with DMARC passing. Mail can authenticate and still fail DMARC when the authenticated identifier does not align with the header From domain.

Those rows are the dangerous ones before reject. They look healthy in a naive ESP dashboard and fall over when policy tightens. Fix alignment first: proper DKIM on the From domain, or SPF that matches how the message is actually sent, within the limits SPF can handle.

Forwarding and mailing lists complicate life. Know which flows you must keep working and test them before the flip.

A sane cutover

Stay on monitor until the report noise is understood. Move to quarantine when you want receivers to start treating failures roughly without the full blast radius. Move to reject when the remaining failures are acceptable losses or already fixed.

Tell the people who send mail. A DNS change without an internal note produces panic tickets that look like an outage.

Keep reading reports after reject. New vendors appear every quarter. DMARC is not a set-and-forget badge; it is an operating habit.

Questions

How long should we stay on quarantine before reject?

Long enough to see weekly and monthly senders in aggregate reports. For many organisations that is several weeks of clean data, not a weekend.

Will reject stop all phishing that uses our brand?

No. It stops unauthorised use of your exact domain in ways receivers honour. Lookalike domains and display-name abuse still need their own controls.

Let's Talk.

Five questions, about two minutes. A named person replies within one working day.

Start a brief