Leaving this course?

Lesson 1 of 5

Why anyone can claim to be you

Email was designed in an era when every machine on the network was assumed to be run by somebody trustworthy. Nothing in the original design verifies who a message is from.

The sender is a claim, not a fact. The address a recipient sees is simply text the sending system chose to put there. There is no more verification in it than in the return address written on an envelope.

There are two sender addresses, and this matters. One is used by the mail servers to route and to return failures. The other is the one displayed to the person reading the message. They can be entirely different, which is both legitimate — a mailing service sends on your behalf — and the gap most impersonation lives in.

What an attacker gains from your domain. An invoice appearing to come from your finance team. A request to change bank details, apparently from a director. A password reset notice to your own staff. The trust is yours; the message is not.

Who is harmed. Usually not you directly. Your customers, suppliers and staff are the ones deceived, which is why this is worth fixing even when your own mail is working perfectly. A domain that can be impersonated is a liability to everybody who trusts it.

The three records are a retrofit. SPF, DKIM and DMARC were added to the existing system rather than designed into it. Each answers one narrow question, and only together do they let a receiving server decide confidently whether a message really came from your domain.

And they are published by you, for others to use. They live in your domain's public records. You are not configuring your own mail server to be suspicious; you are telling every receiving server in the world how to tell your real mail from a forgery.

Lesson 2 of 5

What the three records do

Each one answers a different question, and the common mistake is thinking any of them is sufficient alone.

SPF says which servers may send for your domain. A list, published in your domain records, of the mail servers authorised to send on your behalf. A receiving server checks whether the message arrived from one of them. Its weakness is that it checks the routing address rather than the one the reader sees, and it breaks when mail is forwarded, because the forwarding server is not on your list.

DKIM proves the message was not altered and came from the domain. Your mail server signs outgoing messages with a private key; the matching public key sits in your domain records. The receiver verifies the signature. It survives forwarding, since the signature travels with the message, and it says nothing about which server sent it.

DMARC ties them to the address the reader sees. This is the part that makes the other two useful. It requires that a passing SPF or DKIM result belongs to the same domain as the visible from address — called alignment — and it tells receivers what to do when nothing aligns: do nothing, quarantine, or reject.

Alignment is the whole point. An attacker can trivially pass SPF for their own domain while displaying yours. DMARC is what notices that the passing check and the visible name are different, which is why a domain with SPF and DKIM but no DMARC is still impersonable.

And DMARC asks for reports. Receiving servers send back summaries of what they saw claiming to be you. That feedback is how the rollout in the next lesson is done safely, and it is the only view most organisations ever get of who is forging their domain.

Lesson 3 of 5

Rolling it out without losing mail

The risk in this work is not technical failure. It is turning on enforcement and silently rejecting your own legitimate mail, which is why the sequence matters more than the records.

Find everything that sends as you, first. Marketing platforms, the invoicing system, the recruitment tool, the help desk, the payroll provider, the service that sends password resets, and the server somebody set up four years ago. Organisations consistently discover more senders than they expected, and the ones they forget are the ones that break.

Start DMARC at no action, and ask for reports. Publish a policy that tells receivers to do nothing but report. Nothing changes for any message, and the reports begin arriving. This step is free and it is the one most often skipped.

Read the reports for a few weeks. Identify every source sending as you. Sort them into three piles: ours and authorised, ours and not yet authorised, and not ours. The second pile is the work.

Fix the authorised senders. Add them to SPF, or better, have them sign with DKIM using your domain. Keep the SPF record within the lookup limit the standard sets, which is easy to exceed when many services are added, and which causes checks to fail outright.

Then move in stages. Quarantine a small percentage first, raise it, then reject a small percentage, then raise that. Each step is reversible, and the reports tell you what happened. Moving straight to reject is the error that causes somebody's invoices to vanish.

Watch for the forwarding cases. Mailing lists and automatic forwarding can break SPF legitimately. DKIM usually survives, which is the practical reason to sign rather than rely on SPF alone.

And do it for the domains you do not use. A domain registered and never used for mail should publish records saying so, because an unused domain that sends nothing is a perfect one to forge.

Lesson 4 of 5

What the reports tell you

The feedback is the part people publish and then ignore, which is a shame, because it is the only window most organisations have into their own impersonation.

Aggregate reports are the useful ones. Daily summaries from receiving providers: how many messages claimed to be from your domain, from which network addresses, and whether they passed. They contain no message content and no personal data, so there is no privacy reason to avoid them.

What you learn from them. Which of your own systems are failing checks and why. Which services are sending as you that nobody told you about — usually a department that signed up for something. And the volume of mail claiming to be you from infrastructure that has nothing to do with you, which is the impersonation itself.

Read them as trends. One day's report is noise. The useful questions are whether your authorised volume all passes, whether an unauthorised source is persistent or one-off, and whether a spike coincides with something.

Use a tool. The raw reports are machine-readable files from dozens of providers. Nobody reads them by hand beyond the first week, and a parsing service is the difference between a programme that continues and one that stops after a month.

There is a second kind, and it is different. Failure reports contain parts of individual messages that failed. They are more sensitive, many providers do not send them, and they are not necessary for a successful rollout.

Keep reading after you reach enforcement. This is the step almost everybody skips. A new supplier, a new platform, a renewed certificate, a changed provider — each can break alignment months later, and the reports are what tell you before the complaints do.

And share what you learn. A persistent forging campaign against your domain is worth telling your customers about, and in some cases your national incident response team.

Lesson 5 of 5

What they cannot stop

These records solve one problem completely and leave adjacent ones untouched. Mistaking the scope is how organisations end up believing they have solved phishing.

They stop forgery of your exact domain. That is the win, and it is a real one: nobody can convincingly send mail that appears to come from your actual domain to a receiver that honours the policy.

They do nothing about lookalike domains. A message from a domain one character different, or with a word added, passes its own checks perfectly, because it is a real domain the attacker controls. This is now the common form of business email compromise, precisely because the easy forgery was closed off.

They do not help when a real account is compromised. Mail sent from a genuine mailbox by somebody who stole the password is authentic by every measure these records use. It will pass. This is the attack that matters most for payment fraud.

They say nothing about content. A perfectly authenticated message can carry a fraudulent invoice. Authentication answers who sent it, not whether it is honest.

And they only work where the receiver checks. Major providers do, and smaller or older systems may not. You cannot rely on every recipient enforcing your policy.

So what else is needed. Register and monitor the obvious lookalikes. Flag external mail visibly so an internal-looking message from outside is recognisable. And above all, verify payment and bank detail changes through a channel that is not email, by a number already on file — because that control works against every variant above, and no record published in a domain does.

Knowledge check

Ten questions

Answer all ten questions, then check your answers. You need 9 out of 10 to pass and receive a certificate. If you score less, you will see which answers were right and wrong, and then go through the course again before you retake the check. Your answers, progress and times are kept only in this browser.

Standards and frameworks covered

The published frameworks this course teaches toward. Saolix is not affiliated with, endorsed by or accredited by any of these bodies; naming a standard is a citation, not a claim about its author.

Sources

The official documents this course relies on. Laws and guidance change, so check the current version.

  1. Email security and anti-spoofing guidance · National Cyber Security Centre, United Kingdom
  2. RFC 7208: Sender Policy Framework (SPF) · Internet Engineering Task Force
  3. RFC 6376: DomainKeys Identified Mail (DKIM) Signatures · Internet Engineering Task Force
  4. RFC 7489: Domain-based Message Authentication, Reporting and Conformance (DMARC) · Internet Engineering Task Force
  5. Enhance email and messaging security · Cybersecurity and Infrastructure Security Agency, United States
  6. Business email compromise · Internet Crime Complaint Center, Federal Bureau of Investigation
  7. Computer Misuse and Cybercrimes Act, 2018 (No. 5 of 2018) · Kenya Law
  8. National Kenya Computer Incident Response Team Coordination Centre · Communications Authority of Kenya
  9. ISO/IEC 27001 Information security management systems · International Organization for Standardization
  10. M3AAWG Email Authentication Best Practices · Messaging, Malware and Mobile Anti-Abuse Working Group