How to Find Your DKIM Selector and Confirm Signing Works
The short answer to how to find DKIM selector values is to open the full headers of an email you sent and look for the s= tag in the DKIM-Signature line. That tag is your selector, and RFC 6376, the DKIM standard, explains that receiving servers combine it with the d= domain to locate your public key in DNS. To confirm your mail is signed and passing, check the Authentication-Results header on a received copy for dkim=pass, where header.d is the signing domain, as defined in RFC 8601.
This guide walks through each step in order. You’ll pull the raw headers from a real message and look up the matching key in DNS. You’ll see where platforms like Microsoft 365 keep their selectors, which use a selector1 and selector2 CNAME setup. Then you’ll read the authentication results so you know whether signing actually works.
We’ll also cover what fail, neutral, temperror and permerror mean and what to check first for each. Finally, we’ll look at when it makes sense to have someone else review your setup.
The quick answer: your selector is the s= value in the DKIM-Signature header
Open the full headers of an email your domain sent, find the line that starts with DKIM-Signature, and look for s=. Whatever follows it, up to the next semicolon, is your DKIM selector. Right next to it you’ll usually see d=, which is the signing domain. You need both values to find and check your public key.
Under the DKIM standard, RFC 6376, the s= tag is required in every signature. It works as a label that subdivides the d= domain’s namespace, so one domain can publish several keys side by side. That’s how your email platform, your newsletter tool and your billing system can each sign mail for the same domain without stepping on each other.
How the selector and domain point to your key
When a receiving server checks a signature, it doesn’t guess where your key lives. It combines the selector and the domain from the signature into one DNS name, then looks up the public key there:
- Take the value of s= (the selector).
- Add ._domainkey. after it.
- Add the value of d= (the signing domain) at the end.
The standard uses this example:
| Tag in the signature | Value |
|---|---|
| s= (selector) | brisbane |
| d= (signing domain) | example.com |
| DNS name the verifier queries | brisbane._domainkey.example.com |
If that name returns a valid key, the receiver can test the signature. If it returns nothing or something malformed, verification can’t succeed, no matter how well the sending side is set up.
The selector isn’t a secret
Don’t worry about exposing your selector. It travels in the header of every signed message, so anyone who receives one of your emails can read it. Your security rests on the private key, which stays with your sending system and never appears in DNS or in the message.
Step 1: Open the raw headers of a message you actually sent
Your selector is written into every message your domain signs, so the most reliable place to find it is a real message you sent. Here is how to pull it out.
- Send a real message through your normal sending path. Use the same ESP, CRM or mailbox platform your customers actually receive mail from. Send it to an outside mailbox you control, such as a personal Gmail or Outlook.com address. A test sent from a different tool may be signed differently, or not signed at all.
- Open the original or raw message source. Most mail clients have an option called “Show original,” “View source” or “View raw message.” This shows the full header block that the normal inbox view hides.
- Find each DKIM-Signature header. Search the source for “DKIM-Signature.” Each one is a list of tag=value pairs separated by semicolons.
- Write down the d= and s= values. The d= tag is the signing domain. The s= tag is the selector.
Those two values together tell you exactly where the public key lives. Under RFC 6376, the DKIM standard, the s= tag is required. Receiving servers combine it with d= to look up the key at selector._domainkey.domain. For example, a signature with s=brisbane and d=example.com points to brisbane._domainkey.example.com. That is the name you will look up in Step 2.
If you see more than one signature
It is common for a message to carry several DKIM signatures. For example, it might have one from your ESP’s own domain and one from your domain. Record d= and s= for each, and note which one uses your domain in d=. Order matters too. Google’s DKIM troubleshooting guide says Gmail checks only the first 5 DKIM signatures in Authentication-Results, in the order they appear. Make sure the signature for your domain is among them.
If you want a refresher on what SPF, DKIM and DMARC each actually do, read that first. It explains why the d= domain matters so much later on.

Step 2: Run a DKIM record lookup in DNS
With the selector (s=) and signing domain (d=) from your header, you can check whether the matching public key is actually published. Receivers combine those two values into one DNS name: the selector, then _domainkey, then the domain. The DKIM standard uses the selector “brisbane” and the domain example.com as its example, which gives brisbane._domainkey.example.com as the name a verifier queries.
Query the TXT record at that name using whichever method you prefer:
- dig (macOS, Linux): dig TXT brisbane._domainkey.example.com +short
- nslookup (Windows): nslookup -type=TXT brisbane._domainkey.example.com
- A web lookup tool: enter the selector and domain. You can find options on our free deliverability tools and resources page.
Don’t try to guess selectors or expect DNS to list them for you. DNS has no way to show every selector a domain uses, and the same RFC notes that some domains deliberately choose selector names that can’t be harvested. A real message header is the only reliable way to know which selector to look up.
What a healthy record looks like
- v=DKIM1 at the start, identifying it as a DKIM key record.
- p= followed by a long string of characters. This is the public key.
- An empty p= means the key has been revoked. Mail signed with that selector will not verify, even though a record exists.
Long keys are often split across several quoted strings in the TXT record. Resolvers join those strings back together. If you copied the key into DNS by hand, compare it with the value your provider gave you, character for character, and make sure nothing at the end was cut off. A truncated key looks almost right but will not verify.
If the record is correct and verification still fails, the problem may not be DNS at all. Google notes that a “body hash did not verify” result means the message was changed in transit. Step 3 shows where to see that result.
Where to find the selector in Microsoft 365 and other platforms
If you manage the sending platform, you can usually find the selector in its admin settings. Microsoft 365 is a good example because it works a little differently from most platforms.
Microsoft 365: two selectors, published as CNAMEs
For custom domains, Microsoft 365 uses two selectors, selector1 and selector2. Instead of publishing a public key in a TXT record, you publish two CNAME records that point to keys Microsoft hosts for you. According to Microsoft’s DKIM configuration documentation, the CNAME targets are specific to your tenant, and the dots in your domain are replaced with dashes. Because the exact value varies by tenant, copy the targets straight from the Defender portal or PowerShell rather than building them by hand or reusing an example from a tutorial.
| Hostname | Record type | Value |
|---|---|---|
| selector1._domainkey | CNAME | Copied from Defender portal or PowerShell |
| selector2._domainkey | CNAME | Copied from Defender portal or PowerShell |
Microsoft’s documentation lists a few common mistakes to check for:
- Publishing only selector1. Without selector2, key rotation breaks. Publish both records.
- Using TXT instead of CNAME. Microsoft hosts the keys, so these records need to be CNAMEs.
- Typing the full domain into the hostname field. Many DNS providers add your domain automatically. If you enter it yourself, you end up with a doubled name that verifiers can’t find.
Other platforms
Most email service providers show the selector in their DKIM or domain-authentication settings, often as the first part of the hostname they ask you to publish. If what you see in an admin screen doesn’t match what you expected, go with the s= value in the header of a message you actually sent. That value shows which key receivers are using.
Step 3: Confirm signing works by reading Authentication-Results
Open the raw headers of a message that has arrived in a real inbox, find the Authentication-Results header the receiving server added, and look for dkim=pass. Then check that the signing domain is yours and the selector matches the one you found in Step 1.
The DKIM-Signature header shows what your server claimed. Authentication-Results shows what the receiver concluded after checking it. Under the RFC 8601 definitions of Authentication-Results, a DKIM pass means the message was signed, the receiver found the signature acceptable, and the signature passed verification. The same standard says that header.d reports the d= tag from the signature, so it tells you which domain actually signed the message.
A typical result line looks something like this:
dkim=pass header.d=yourdomain.com header.s=yourselector
As you read it, check three things:
- The result says pass. Anything else needs attention. The next section covers the other results.
- header.d is your domain. A pass for your email provider’s shared sending domain proves that the provider signed the message. It does not prove that your own domain did. Look for a pass that names your domain.
- header.s matches your selector. If the receiver shows header.s, it should match the s= value you found in Step 1. A different selector can mean another service signed the message, or an old key is still in use.
If the result is dkim=none, RFC 8601 treats that as meaning the message wasn’t signed. In that case, go back to your sending platform and confirm that DKIM signing is turned on for this domain.
Some messages carry several signatures and show several dkim= results. Google’s DKIM troubleshooting guide notes that Gmail checks only the first five DKIM signatures, in order. Make sure the signature for your domain isn’t buried behind others.
Checking one message at a time works for a spot check. To see results across all your mail streams and every receiver, start reading your DMARC aggregate reports.
DKIM signature fail, neutral, temperror, permerror: what each means and what to check first
Each result points to a different kind of problem, so the fix starts with reading the exact word in Authentication-Results. The definitions below come from RFC 8601, the standard for the Authentication-Results header.
fail
The signature was accepted for checking but didn’t verify. What to check first:
- If Gmail reports “body hash did not verify,” the message was changed after it was signed. Google’s DKIM troubleshooting guide notes that an outbound gateway adding a footer is a common cause.
- Look for anything that touches mail after signing, such as disclaimer tools, security gateways or link rewriters.
neutral
The message was signed, but the signature had syntax errors or couldn’t be processed. Check the DKIM-Signature header for malformed or missing tags, and confirm your sending platform is generating it correctly.
temperror
Something likely temporary went wrong, such as the receiver being unable to fetch your public key. Trying again later may produce a final result. If it keeps happening:
- Confirm your DNS host is answering reliably.
- Check the TTL on your DKIM records. Microsoft links very low TTLs to intermittent temperror and recommends setting them to at least 3600 seconds.
permerror
A permanent problem, such as a required header field being missing, stopped verification. Retrying won’t help; the signature or the record itself has to be fixed.
Microsoft 365 setup mistakes
Two errors from that same Microsoft guide come up often, and both show up as CnameMissing:
- Entering the full domain in the hostname field when your DNS host adds the domain automatically.
- Publishing a TXT record instead of a CNAME.
If you need to rebuild or correct your records, our step-by-step SPF, DKIM and DMARC setup for high-volume senders walks through each one in order.

When it makes sense to get a second pair of eyes
If you send from one platform and your headers show a DKIM pass for your own domain, you may not need outside help at all. Tracing gets harder when mail leaves your domain from several places at once.
Senders who use several platforms (for example a workspace mailbox, a newsletter tool, a CRM and a billing system) often have several selectors. Each one needs its own published key in DNS and its own passing result in Authentication-Results. One healthy selector tells you nothing about the others. Your everyday mailbox can test fine while invoices or campaigns quietly fail.
A few situations make this harder to untangle on your own:
- Several signatures on one message. A message can carry more than one DKIM signature. Google notes that Gmail checks only the first 5 DKIM signatures, in order, so the way your platforms stack signatures can affect what gets evaluated.
- Gateways and relays. An outbound gateway that adds footers or rewrites content can cause failures even when the published key is correct.
- Key rotation. During a rotation, old and new selectors exist side by side. It’s easy to remove a record that one platform still relies on.
If results stay inconsistent after working through the checks above, an email deliverability audit can map every sending source to its selector and DNS record. That shows you which platform signs with which key and where the gaps are.
If you’d like a hand with that, you’re welcome to book a free consult. We’ll look at what you’re seeing in your headers and talk through sensible options. There’s no obligation and we won’t promise specific outcomes. The goal is simply a clearer picture of how your mail is being signed.
Keep researching on Digital Bulldogs
Frequently asked questions
Can I find a domain’s DKIM selector without having one of its emails?
Not reliably. DNS has no index of selectors. A verifier only finds the key because the signature names both the domain (d=) and the selector (s=), then queries selector._domainkey.domain. The DKIM standard, RFC 6376, also notes that senders may choose selectors that are deliberately hard to guess. You can try common names, but a miss doesn’t prove there’s no key.
The dependable route is to get a message that was actually sent from the domain, or ask whoever manages its email. If you do find a record and its p= value is empty, that key has been revoked and isn’t the one in use.
Can one email have more than one DKIM signature?
Yes, and it’s common. Your own domain might sign a message, and the sending platform might add its own signature. Each one appears as a separate DKIM-Signature header with its own d= and s= values. Order matters at Gmail, though. According to Google’s DKIM troubleshooting guide, Gmail checks only the first 5 DKIM signatures, in order, and reports them in Authentication-Results. If the signature you care about sits further down, Gmail may never evaluate it.
Why does Microsoft 365 use selector1 and selector2?
Two selectors let Microsoft rotate keys without interrupting signing. One key stays active while the other is ready to take over. Each selector is published as a CNAME that points to a tenant-specific target, and Microsoft’s DKIM configuration guide recommends copying the exact values from the Defender portal or PowerShell rather than typing them by hand.
Microsoft lists several frequent mistakes: leaving out selector2, publishing TXT records instead of CNAMEs, putting the full domain in the hostname field, and setting a TTL that is too low. Microsoft advises a TTL of at least 3600 seconds.
What does dkim=temperror mean, and will it go away on its own?
A temperror means the receiver hit a temporary problem while verifying, often a DNS lookup that didn’t complete. Under RFC 8601, it signals that a later attempt might succeed. That’s different from a fail, where the signature was checked and didn’t verify.
A one-off temperror can clear up without any action. If you see it repeatedly, look at your DNS. Check whether your DKIM records resolve consistently, and for Microsoft 365, confirm the TTL meets the guidance above.
What causes ‘body hash did not verify’?
The message body changed after it was signed. DKIM stores a hash of the body in the signature. If the receiver’s hash doesn’t match, the DKIM specification tells the verifier to ignore that signature and report “body hash did not verify.”
The usual cause is a system between signing and delivery that edits the content. Outbound gateways that add legal footers or disclaimers are a frequent example, and so are mailing lists that add their own text. The fix is to have that system leave messages untouched or to sign after it runs.
Sources
- RFC 6376: DomainKeys Identified Mail (DKIM) Signatures · IETF
- RFC 8601: Message Header Field for Indicating Message Authentication Status · IETF
- Troubleshoot DKIM issues · Google Workspace Help
- How to use DKIM for email in your custom domain · Microsoft Learn
- RFC 6376: DomainKeys Identified Mail (DKIM) Signatures – Pike Programming Language
This article is general information, not legal advice. Laws and deadlines vary by state and change over time; check the linked sources or speak with a licensed attorney about your situation.

