How to Authenticate Your Sending Domain, in Plain Terms
How to authenticate your sending domain: what SPF, DKIM and DMARC each prove, why a domain you own matters, and what to verify before any nurture mail goes out.
To authenticate your sending domain, publish all three records, SPF, DKIM and DMARC, on a domain you own, and verify that they work before you send nurture mail at any volume. Mailbox providers trust a message partly by checking that it really came from who it claims to be from, and these three records are how you let them check.
This post stays at the level of what each record is for and what you need to have in place. It does not walk through the entries themselves, because those depend on your own setup. What it gives you is the reasoning, so that you know what you are asking for and can tell whether it has been done.
How to authenticate your sending domain: what you are proving
Anyone can write a message that claims to come from your name. Mailbox providers know that, so they do not take a sender's word for it. They check, and the result of the check is one of the things that decides whether your message goes to the inbox, goes to spam or is never accepted at all.
Authentication is therefore a form of proof. It tells the provider that this message was sent by a server you allowed, that it was not altered on the way and that you have a stated policy for messages that fail those checks. A sender who cannot give that proof is treated with suspicion, however good the email is.
This is why authentication belongs at the front of any nurture plan. Nothing you do later, however thoughtful, can rescue mail that providers do not trust. The wider picture of why mail fails is in email deliverability for founders.
The three records and what each one does
| Record | What it does | The question it answers for the provider |
|---|---|---|
| SPF | Lists the servers allowed to send for your domain | Is this message coming from a server the domain owner approved? |
| DKIM | Signs each message so it provably was not altered | Is this message exactly what the sender sent? |
| DMARC | Tells providers what to do when a message fails those checks | What does the owner want me to do with a message that fails? |
Think of them as three parts of one argument. The first says who may send. The second says the message is genuine. The third says what to do with anything that fails the first two. Each is weaker without the others, which is why the rule is to have all three, not to choose the one that seems easiest.
Why all three, and not just one
A common instinct is to do the first and assume the job is finished. That leaves gaps. A list of approved servers says nothing about whether the content was changed, and a signature says nothing about what should happen to forgeries. The policy record exists precisely so that the owner, not the provider, decides what a failure means. You need all three in place on a domain you own before you send nurture at any volume.
It has to be a domain you own
The records are published on a domain, so the domain has to be one you control. That sounds obvious, and it is the point on which the most common failure turns. An unauthenticated or shared domain is the most common reason careful sequences land in spam.
A shared domain means your reputation is tied to the behaviour of everyone else using it. You can do everything right and still be judged by the company you are keeping. A domain you own gives you a reputation that is yours alone, built from your own sending, and it gives you the ability to publish the records that prove it.
Decide whether marketing gets its own subdomain
Many teams send marketing from a subdomain. The reason is separation: if a marketing reputation problem occurs, it cannot contaminate the domain that carries invoices and password resets. Those are the messages people truly need, so it makes sense to keep them away from the risks that come with bulk sending.
This is a decision to make once, early, and write down. The checklist version is a single line: I have decided whether marketing mail uses its own subdomain. Settle it before you send at volume, so that the domain you authenticate is the domain you actually send from.
Verify, do not assume
Publishing a record and having it work are not the same thing. The standard to meet is that SPF, DKIM and DMARC are all published and verified on the domain. The word "verified" is doing real work in that sentence.
The reason is the silent nature of the failure. When authentication is missing or wrong, nothing visibly breaks. Your sequence runs, your dashboard shows sends and a meaningful share of the mail ends up in spam or is never accepted. A sender who only checks that the sequence is sending will not discover the problem until the results disappoint, and by then the cause looks like weak writing.
| State | What you see | What is really happening |
|---|---|---|
| Authenticated and verified | Sends, and mail reaching inboxes | Providers can confirm who you are |
| Missing or incorrect | Sends, exactly as above | Mail is treated with suspicion and may land in spam |
Look at the two rows. From the dashboard they are identical. Only a check of the authentication itself tells them apart, which is why verification is part of the job.
Who should check, and how often
Everything in a nurture programme can be done with an email tool, a spreadsheet and a careful person who checks the exits, the suppression list and the authentication every week. That person does not need to be technical in any deep sense. They need to own the question: are the records still published and still verified?
It is easy to forget authentication, because nobody revisits it once it is working. The authentication nobody remembers setting up is one of the quiet failure points in any programme. Put it in the monthly review so that it is looked at on a calendar and not only when results fall.
Authentication is one part of the sending identity
The records prove who you are to the provider. The person reading the message also needs to see a sender they can trust. That is why the same checklist asks that mail comes from a named person at an address that accepts replies. A no-reply address tells the reader you do not want to hear from them, and replies are among the strongest buying signals you will get.
The two work together. Authentication gets the message accepted. A real name and a reply address make it feel like a message from a person. Both are explained in the context of the whole sequence in how to build an email nurture sequence.
Questions to put to whoever sets it up
You may not be the person who publishes the records, and you do not have to be. You do have to ask the right questions and expect plain answers.
- Which domain are we sending from, and do we own it?
- Are SPF, DKIM and DMARC all published on it, not just one or two?
- Has each been verified, and by what check?
- Does marketing mail use its own subdomain, and was that a deliberate choice?
- Who looks at this again, and when?
A person who can answer all five has the matter under control. A person who answers "it should be fine" has given you a reason to check. Sending from an unauthenticated or shared domain is among the most damaging mistakes in the whole discipline, precisely because careful work lands in spam behind a dashboard full of sends.
Your pre-send authentication checklist
- I send from a domain I own, not a shared one.
- SPF, DKIM and DMARC are all published and verified on that domain.
- Mail comes from a named person at an address that accepts replies.
- I have decided whether marketing mail uses its own subdomain.
If any line is unticked, fix it before the first sequence goes live. The cost of doing it first is a short afternoon. The cost of discovering it later is a programme that looked healthy while it was not.
Frequently asked questions
Do I need all three records?
Yes. SPF, DKIM and DMARC do different jobs, and you need all three in place before you send nurture mail at any volume.
Can I send from a shared domain?
It is better not to. An unauthenticated or shared domain is the most common reason careful sequences land in spam.
Why would I use a subdomain for marketing?
So that a marketing reputation problem cannot contaminate the domain that carries invoices and password resets.
How would I know authentication is broken?
You often would not, because the failure is silent: the sequence keeps sending and the dashboard keeps counting. That is why you verify the records and recheck them on a schedule.
The full method, with the checklists, is in The Email Nurture Playbook, a free 34-page playbook.