A customer forwards you an email "from" your accounts department, asking them to pay an invoice into a new account. Nobody on your team sent it. Meanwhile, your real invoices keep ending up in your customers' junk folders, and nobody can explain why.
Both problems usually have the same root cause: your domain doesn't tell the rest of the internet which email is really yours. Three DNS records fix that: SPF, DKIM and DMARC. They cost nothing to publish, and together they are the standard defence against criminals sending email that appears to come from your exact domain.
The Centre for Cybersecurity Belgium (CCB) recommends all three on Safeonweb@Work, both to protect against phishing and spoofing and to reduce the chance that your own mail is marked as spam. Here is what each one does, and how to switch them on without blocking your own email.
What the three records do
Think of an email as a letter. Anyone can write any return address on the envelope. These records give the receiving mail server a way to check it.
| Record | What it says | What it proves |
|---|---|---|
| SPF | "These servers may send email for my domain" | The message came from an authorised server |
| DKIM | "My messages carry this digital signature" | The message really came from the domain and wasn't altered |
| DMARC | "If a message fails these checks, do this, and send me a report" | The address your recipient sees matches what was checked |
SPF: the list of approved senders
A Sender Policy Framework (SPF) record is a line of text in your domain's DNS that lists the services allowed to send email for you. For a Microsoft 365 business, it usually starts as v=spf1 include:spf.protection.outlook.com -all.
Two rules trip people up, according to Microsoft. You may have only one SPF record per domain, so a second one added by your web developer breaks both. And the record may trigger no more than 10 DNS lookups. Every include: for a newsletter tool, CRM or booking system counts, so a long record can silently fail.
DKIM: the signature
DomainKeys Identified Mail (DKIM) adds a digital signature to each outgoing message, and publishes the matching public key in your DNS. The receiving server checks the signature and knows the message wasn't altered and was signed by your domain (Safeonweb@Work).
Here is the detail many Microsoft 365 tenants miss. Microsoft automatically signs mail with your initial onmicrosoft.com domain, but DKIM signing with your own custom domain has to be configured: you publish two CNAME records (selector1 and selector2) and then switch signing on in the Defender portal. Until you do, the signature doesn't match the domain your customers see, which matters for DMARC.
DMARC: the policy that ties it together
Domain-based Message Authentication, Reporting and Conformance (DMARC) checks that the domain in the visible From address lines up with the domain that passed SPF or DKIM. Microsoft explains that a message passes DMARC if at least one of the two checks passes and aligns with the From address.
The DMARC record, published at _dmarc.yourdomain.be, then tells receivers what to do with mail that fails. There are three policies:
p=none: take no action, just report. Useful for learning, but it doesn't stop spoofing.p=quarantine: deliver failing mail to the junk folder.p=reject: refuse failing mail outright.
DMARC also lets you request aggregate reports (the rua= tag): daily summaries from receiving providers showing which servers sent mail using your domain, and whether it passed. Those reports are how you find the forgotten systems that send on your behalf.
Why this is no longer optional
The large mailbox providers now enforce these records. Google requires every sender to Gmail to use SPF or DKIM, and anyone sending more than 5,000 messages a day must have SPF, DKIM and DMARC, with the From domain aligned. Since May 2025, Microsoft's Outlook.com has applied similar rules to high-volume senders: SPF and DKIM must pass and DMARC must be at least p=none, or the message is rejected.
Most SMBs don't send 5,000 emails a day. But a newsletter to your whole customer base can. And whatever your volume, the CCB notes that these records reduce the likelihood that your email is marked as spam.
The standard itself has also matured. In May 2026 the IETF published DMARC as a full standards-track specification, RFC 9989. Existing DMARC records keep working under the new version, so this is not a reason to rush.
Going from p=none to p=reject safely
The risk with DMARC isn't the criminals: it's blocking your own legitimate email. The scanner that emails PDFs, the accounting package that sends invoices, the newsletter platform and the website contact form may all send "as" your domain. If they aren't covered by SPF or DKIM, a strict policy will block them too.
That's why both the CCB and Microsoft recommend a stepwise approach. The CCB's DMARC implementation guide describes starting at none, moving to quarantine once you trust the configuration, then to reject once every legitimate source is set up correctly. Its Cyberguide advises monitoring for at least 30 days before tightening the policy.
The goal is
p=reject. A DMARC record left atp=nonefor years gives you reports, but it still lets anyone send email as your domain.
What to do:
- Make an inventory of everything that sends email with your domain in the From address. The DMARC reports will show you the ones you forgot.
- For each third-party service, configure it to DKIM-sign with your own domain. Many newsletter and CRM tools support this. It is more reliable than adding yet another
include:to SPF. - Read the reports, or have someone read them for you. The raw files are XML and not meant for humans: a DMARC reporting tool turns them into a readable list of senders.
Don't forget your other domains
Many businesses own several domains: the old company name, the .com next to the .be, a campaign domain. Criminals like unused domains because nobody watches them.
For any domain that never sends email, Microsoft recommends an SPF record of v=spf1 -all (nobody may send) and a DMARC record of v=DMARC1; p=reject;. That takes five minutes per domain and closes the door completely.
What DMARC doesn't do
DMARC protects your exact domain. It does nothing against a lookalike such as go-secvre.be, and nothing against a genuine mailbox that has been taken over. Those are the tactics behind most invoice fraud, and they need different controls: impersonation protection, alerts on suspicious inbox rules, and a payment process that never trusts email alone.
Where to start
- Check your current records. Ask whoever manages your DNS, or use a free online DMARC checker, to see whether SPF, DKIM and DMARC exist for every domain.
- Turn on DKIM for your custom domain in Microsoft 365, and fix your SPF record so there is exactly one, under 10 lookups.
- Publish DMARC at
p=nonewith a reporting address, and collect reports for at least a month. - Bring every legitimate sender into line, then move to
p=quarantineand finallyp=reject. - Lock down every domain that doesn't send email with
v=spf1 -allandp=reject.
The takeaway
SPF, DKIM and DMARC are three short lines of text that tell the world which email is genuinely yours. Set them up step by step, use the reports to find every system that sends on your behalf, and don't stop at p=none. With a month of reports and some patient clean-up, most small businesses can get to p=reject. From then on, mail servers that check DMARC will refuse email that pretends to come from your domain.