DMARC for Parked Domains: SPF, Null MX and a Lockdown Checklist
Every domain you own that never sends email needs three DNS records: an SPF record that reads v=spf1 -all, a DMARC record set to p=reject, and a null MX record showing the domain accepts no mail. That trio is the core of DMARC for parked domains. It tells receiving mail servers that any message claiming to come from that domain is not legitimate, and that the domain has no inbox to reply to. The M3AAWG best common practices for protecting parked domains recommend SPF -all and DMARC p=reject for exactly this kind of domain.
This matters because a quiet domain is not a safe domain. When these records are missing, nothing tells receivers to distrust mail that uses your domain name. That leaves room for someone to put your brand in the From line. For direct-response marketers, affiliates and agencies, that kind of abuse can drag on the reputation you rely on for inbox placement from your real sending domains.
Below, we explain why unprotected domains put your main brand at risk. We then cover each record in turn: the SPF syntax defined in RFC 7208, a DMARC reject policy with reporting under the current standard, RFC 9989, and the null MX record from RFC 7505. We also look at the optional empty DKIM record. Finally, we show you how to inventory every domain you own and give you a checklist you can work through one domain at a time.
The short answer: three records for a domain that sends no email
If a domain you own will never send email, publish three DNS records that say so: an SPF record that authorizes no senders, a DMARC record that tells receivers to reject anything claiming to come from the domain, and a null MX record that says the domain accepts no mail. Together they turn a blank spot in DNS into a clear “nothing legitimate comes from here.”
A parked domain is one you registered with no plan to send or receive email, or one you use only for a website. The M3AAWG best common practices for protecting parked domains recommend that domains like these publish records stating that they don’t send mail, rather than leaving receivers to guess.
| Record | Where it goes | Value | What it tells receivers |
|---|---|---|---|
| SPF (TXT) | Domain root | v=spf1 -all | No server is authorized to send mail for this domain, per RFC 7208 |
| DMARC (TXT) | _dmarc.yourdomain | v=DMARC1; p=reject plus a reporting address | Reject mail that fails authentication and send you reports about it |
| Null MX | Domain root | Preference 0, target . | This domain accepts no email, as defined in RFC 7505 |
You can also add an empty DKIM wildcard record. It’s optional, and we cover it later in this guide.
If you set up DMARC records under the older specification, RFC 7489, you don’t need to rewrite them. RFC 9989 replaces RFC 7489 as the current DMARC standard, and records that begin with v=DMARC1 remain valid under it.
Why unprotected parked domains put your main brand at risk
A parked domain with no email records is an easy target. It looks legitimate, it carries your brand name, and nothing in DNS tells receiving mail servers that it should never send a message. The M3AAWG best common practices for protecting parked domains warn that when you don’t publish records saying a parked domain sends and receives no email, you leave openings for third parties to abuse it.
Lookalike and typo domains are the most exposed
M3AAWG points out that domains resembling well-known domains face particular risk. That describes many of the domains brands register defensively: common misspellings, alternate extensions and close variations of the main name. You may have bought them to keep them out of someone else’s hands. Left without email records, they are still the domains an impersonator would most like to use.
No record means no instruction to reject
When a forged message arrives claiming to come from your parked domain, the receiving server looks for guidance from the owner. If there is no SPF record and no DMARC record, it finds none. Nothing tells it the mail is unauthorized, and nothing asks it to reject the message. The receiver has to rely on its own filtering.
Why this is a deliverability problem, not only a security one
Recipients don’t always distinguish between your main domain and a close variant. If phishing or spam goes out under a name tied to your brand, people can learn to distrust anything that looks like it came from you. That doubt can carry over to your legitimate campaigns. Locking down parked domains therefore protects the mail stream you depend on, as well as your security posture.
Prevention and cleanup are separate jobs. The records in the next sections stop future abuse, but they don’t undo damage that has already been done. If impersonation has already hurt how your mail is received, repairing a domain’s reputation after impersonation is its own process with its own steps.
SPF for a non-sending domain: v=spf1 -all
If a domain never sends email, publish this single TXT record at the domain’s root:
v=spf1 -all
RFC 7208, the SPF standard, gives this exact record as the way for a domain to say it sends no mail. The record lists no mechanisms (no IP ranges, no include: entries), so no server is ever authorized. Every message claiming to come from the domain gets an SPF fail result.
Why it has to be “-all”
The “-all” at the end is what makes the answer definite. RFC 7208 explains that receivers can only conclude “not authorized” when a record ends in “-all” or redirects to a record that does. Weaker endings don’t give receivers that clear answer:
- ~all (softfail) means “probably not authorized.” Many receivers treat that as a reason for suspicion, not rejection.
- ?all (neutral) makes no claim either way.
A domain that sends nothing has no legitimate mail to protect, so it doesn’t need the softer setting. The M3AAWG parked domains best practices also recommend “v=spf1 -all” for domains that never send.
Common mistakes to avoid
- Publishing two SPF records. Use exactly one v=spf1 TXT record per name. Under RFC 7208, more than one causes a permanent error, which undoes the protection you meant to add. This often happens when an old registrar default sits next to the new record, so delete the old one.
- Leaving out other hostnames. SPF is checked against the specific name used in the message, not the whole zone. If hostnames like www or shop have A records, consider giving each one the same “v=spf1 -all” record.
- Copying a sending domain’s record. Pasting your main domain’s include: list onto a parked domain authorizes servers that should never send for it.
If you manage many domains, it helps to handle DNS authentication setup for your sending and non-sending domains as one project. That way each parked name gets the same clean record set.

DMARC p=reject with reporting for parked domains
A domain that never sends email should publish a DMARC record with p=reject, and it can start there on day one. The M3AAWG best common practices for parked domains recommend exactly this. The usual slow ramp from p=none to quarantine to reject exists to protect legitimate mail while you find every sending source. A parked domain has no legitimate mail, so there is nothing to protect and no reason to wait.
A minimal record, published at _dmarc.yourparkeddomain.com, looks like this:
v=DMARC1; p=reject; rua=mailto:dmarc-reports@yourmaindomain.com
Manage every parked domain from one record
If you hold dozens of defensive or lookalike registrations, editing each one by hand gets messy. M3AAWG describes a cleaner approach. You point each parked domain’s _dmarc name by CNAME to a single central parked-domain DMARC record. When you change the policy or the reporting address, you change it once, and every parked domain follows.
Read reports as evidence, not errors
The rua tag tells receivers where to send aggregate reports. On a parked domain, every report entry shows someone trying to send as that domain. That is useful evidence of spoofing attempts and of which networks they come from. It does not mean your setup is broken. With p=reject in place, receivers that honor DMARC should be refusing those messages.
What RFC 9989 changed
RFC 9989 was published in May 2026 and replaces RFC 7489 as the DMARC standard. Existing v=DMARC1 records remain valid. The new standard adds an np tag, which sets policy for subdomains that don’t exist. If you leave np out, receivers fall back to sp, then to p.
For a parked domain, that fallback works in your favor. One p=reject record covers the whole tree unless a subdomain publishes its own record. Policy for live subdomains that do send, set with the sp tag, is covered in our separate subdomain guide.
Null MX (RFC 7505) and the optional empty DKIM record
If a parked domain should never receive email, publish a null MX record. It tells every mail server, in one line, that the domain accepts no mail.
Here is why that matters. When a domain has no MX record, sending servers fall back to its A or AAAA address. A parked domain that still hosts a website or redirect page can therefore look like it accepts mail, even though nobody is reading it. RFC 7505 defines the null MX record to close that gap. It is a single MX record with preference 0 and a target of “.” (a lone dot):
- Record: example.com. MX 0 .
- Rule: A domain that publishes a null MX must not publish any other MX record.
- Receiver behavior: The RFC says receivers should reject mail whose sender domain has a null MX, typically with a 556 or 550 reply code.
That last point gives you a second signal on top of SPF and DMARC. Even a receiver that handles your authentication records loosely has a separate reason to refuse forged mail from the domain.
When to skip it: Only add a null MX if the domain truly must not receive mail. If a domain still takes inbound messages, for example as a forwarding alias to your main inbox, a null MX will break that. Leave those domains out of this step.
The optional empty DKIM key
You can also publish a DKIM record with an empty key at a wildcard selector, such as *._domainkey.example.com TXT “v=DKIM1; p=”. An empty p= means no valid key exists, so no signature can verify. This step is optional. M3AAWG’s parked domain guidance says parked domains shouldn’t publish working DKIM keys. Before you add the wildcard, look for old selectors left over from past senders or platforms, and remove any live keys you find.
DNS timing for parked domains
The same M3AAWG document suggests DNS timing values for parked domains. These include SOA settings measured in hours to a day, rather than the shorter values active domains use. Parked records rarely change, so longer timers reduce lookup load without slowing down any real updates.
How to inventory every domain you own before locking it down
Before you change a single DNS record, list every domain you own and what each one actually does. Publishing lockdown records on a domain that still sends mail will break that mail, so the inventory comes first.
Find every domain
- Export domain lists from every registrar account, DNS host and reseller. Include accounts opened by past staff, acquired companies, agencies and affiliates.
- Add defensive typo and lookalike registrations. The M3AAWG best common practices for protecting parked domains flag lookalike domains as a real abuse risk, so they belong on the list even if nobody has checked them in years.
- Include old product and campaign domains, regional TLDs and domains kept only for redirects.
Record what each domain does today
For each domain, note whether it serves a website only, redirects, receives mail, sends mail (directly or through a vendor), or does nothing. Then check its current MX, SPF and _dmarc records. Look for leftover DKIM selectors that point to an email platform, CRM or other vendor that may still send on the domain’s behalf.
| Category | Typical signs | Next step |
|---|---|---|
| Sending | SPF includes, live DKIM selectors, an active vendor, or a mail path you can’t explain | Leave it out of the parked-domain change and secure it as a sending domain |
| Receive-only | MX points to a working mailbox, with no outbound mail | Keep inbound mail working and lock down sending only |
| Fully parked | No mail in or out. It’s a website, a redirect, or nothing at all | Candidate for the full lockdown |
When in doubt, treat it as sending. Any domain with an unknown or active mail path stays out of the parked-domain change until someone confirms it no longer sends.
Agencies and affiliate networks that manage many client domains can keep this inventory in a shared record that is reviewed on a regular schedule. Note each domain’s owner, registrar, purpose and the date it was last checked. If you’d like a hand with that, we offer support for agencies managing client domain portfolios.

Parked domain lockdown checklist
Work through these steps for each domain that should never send email. Do them in order, because the first step confirms that the later steps won’t block mail you actually need.
- Confirm the domain has no legitimate sender. Check that no ESP, CRM, helpdesk, billing or transactional tool uses the domain in a From or envelope address. If something does, move that tool to an active domain before you lock this one down.
- Publish v=spf1 -all at the root and on every host name that resolves. This is the record RFC 7208 describes for domains that send no mail. Keep exactly one SPF record per name, because a second record causes a permanent error.
- Publish a _dmarc record with p=reject and a rua address you actually monitor. You can also point it by CNAME to a central record. You may add np=reject, the tag RFC 9989 defines for non-existent subdomains. Existing v=DMARC1 records remain valid.
- Publish a single null MX (0 .) if the domain must not receive mail. RFC 7505 says no other MX records may sit alongside it, so delete any old ones.
- Optionally publish an empty-key DKIM wildcard. Also remove any active DKIM selectors that are no longer in use.
- Recheck after every DNS host or registrar change. Records can be lost during a migration. Add each new registration to your inventory as soon as it is created, and lock it down on the same day.
- Review DMARC aggregate reports for spoofing attempts. On a parked domain, any traffic that claims to come from it deserves a closer look.
The M3AAWG parked domains best common practices cover this same approach in more depth, including the central CNAME setup and DNS timing.
If you manage a large portfolio and want a second pair of eyes, book a free consult. We can walk through your inventory with you and explain how our email delivery and authentication services could support your active sending domains as well.
Frequently asked questions
Can a parked domain go straight to DMARC p=reject without starting at p=none?
Yes. The usual p=none phase exists so you can find legitimate mail streams before enforcing a policy. A parked domain has no legitimate mail streams, so there is nothing to discover and nothing to break. The M3AAWG best common practices for parked domains recommend publishing DMARC p=reject for domains that never send. Add a reporting address so you can see anyone trying to use the name.
Will a null MX record break the website on my parked domain?
No. A null MX only describes how the domain handles mail. It does not affect the A or AAAA records that serve a website or a parking page. RFC 7505 defines the null MX as a single MX record with preference 0 and a target of “.”. Its purpose is to stop senders from falling back to your A/AAAA records when they try to deliver mail. Visitors can still load the site.
Does a domain that never sends email need a DKIM record?
It isn’t strictly required. SPF -all and DMARC p=reject already tell receivers that no mail from the domain is legitimate. The same M3AAWG parked-domain guidance treats DKIM as an optional extra layer. You can publish a wildcard DKIM record with an empty public key, such as *._domainkey TXT "v=DKIM1; p=". An empty key signals that no valid signing key exists, so any signature claiming to come from the domain will fail.
What if the domain receives email but never sends it?
Keep your normal MX records and skip the null MX. RFC 7505 says a null MX must not be published alongside other MX records, because it declares that the domain accepts no mail at all. You can still lock down sending. RFC 7208 gives v=spf1 -all as the record for domains that send no mail, and DMARC p=reject works the same way.
Check one thing first: auto-replies, forwarding rules and out-of-office messages count as sending. If any of those are active on the domain, it isn’t truly send-free. If you’re unsure how your domains are set up, you can book a free consult and we’ll review them with you.
Do I need to rewrite my parked-domain DMARC records now that RFC 9989 has replaced RFC 7489?
No. RFC 9989, the current DMARC standard, keeps existing v=DMARC1 records valid, so a working p=reject record can stay as it is. The update adds an optional np tag that sets the policy for non-existent subdomains. If np is absent, receivers fall back to sp and then to p. A parked domain already set to p=reject is therefore covered. Adding np=reject simply makes that intent explicit.
Sources
- M3AAWG Protecting Parked Domains Best Common Practices (updated June 2022) · Messaging, Malware and Mobile Anti-Abuse Working Group (M3AAWG) · 2022-06-01
- RFC 7208: Sender Policy Framework (SPF) for Authorizing Use of Domains in Email, Version 1 · IETF · 2014-04-01
- RFC 7505: A "Null MX" No Service Resource Record for Domains That Accept No Mail · IETF · 2015-06-01
- RFC 9989: Domain-Based Message Authentication, Reporting, and Conformance (DMARC) · RFC Editor / IETF · 2026-05-19

