Choosing the Right Domain for Optimal Deliverability with Amazon SES

As a sender, selecting the right domain for the visible From header of your outbound messages is crucial for optimal deliverability. In this blog post, we will guide you through the process of choosing the best domain to use with Amazon Simple Email Service (SES)

Understanding domain selection and its impact on deliverability

With SES, you can create an identity at the domain level or you can create an email address identity. Both types of verified identities permit SES to use the email address in the From header of your outbound messages. You should only use email address identities for testing purposes, and you should use a domain identity to achieve optimal deliverability.

Choosing the right email domain is important for deliverability for the following reasons:

• The domain carries a connotation to the brand associated with the content and purpose of the message.
• Mail receiving organizations are moving towards domain-based reputational models; away from IP-based reputation.
• Because the email address is a common target for forgery, domain owners are increasingly publishing policies to control who can and cannot use their domains.

The key takeaway from this blog is that you must be aware of the domain owner’s preference when choosing an identity to use with SES. If you do not have a relationship with the domain owner then you should plan on using your own domain for any email you send from SES.

Let’s dive deep into the technical reasons behind these recommendations.

What is DMARC?

Domain-based Message Authentication, Reporting, and Conformance (DMARC) is a domain-based protocol for authenticating outbound email and for controlling how unauthenticated outbound email should be handled by the mail receiving organization. DMARC has been around for over a decade and has been covered by this blog in the past.

• What is DMARC and should you use it?
• Complying with DMARC across multiple accounts using Amazon SES

DMARC permits the owner of an email author’s domain name to enable verification of the domain’s use. Mail receiving organizations can use this information when evaluating handling choices for incoming mail. You, as a sender, authenticate your email using DKIM and SPF.

• DKIM works by applying a cryptographic signature to outbound messages. Mail receiving organizations will use the public key associated with the signing key that was used to verify the signature. The public key is stored in the DNS.
• SPF works by defining the IP addresses permitted to send email as the MAIL FROM domain. The record of IP addresses is stored in the DNS. The MAIL FROM domain is not the same domain as the domain in the From header of messages sent via SES. It is either domain within amazonses.com or it is a custom MAIL FROM domain that is a subdomain of the verified domain identity. Read more about SPF and Amazon SES.

A message passes the domain’s DMARC policy when the evaluation DKIM or SPF indicate that the message is authenticated with an identifier that matches (or is a subdomain of) the domain in the visible From header.

How can I look up the domain’s DMARC policy?

You must be aware of the DMARC policy of the domain in which your SES identities reside. The domain owner may be using DMARC to protect the domain from forgery by unauthenticated sources. If you are the domain owner, you can use this method to confirm your domain’s current DMARC policy.

You can look up the domain’s DMARC policy in the following ways:

• Perform a DNS query of type TXT against the hostname called _dmarc.<domain>. For example, you can use the ‘dig’ or ‘nslookup’ command on your computer, or make the same query using a web-based public DNS resolver, such as https://dns.google/
• Use a 3rd party tool such as:

https://tools.wordtothewise.com/dmarc/
https://mxtoolbox.com/dmarc.aspx
https://dmarcian.com/dmarc-inspector/

The “p” tag in the DMARC record indicates the domain’s policy.

How does the domain’s policy affect how I can use it with SES?

This section will cover each policy scenario and provide guidance to your usage of the domain with SES.

Other considerations

If the domain has a none or quarantine policy, you must be aware that the domain owner may have a plan to migrate to a more restrictive policy without consulting with you. This will affect your deliverability in the form of low inboxing/open rates, or high bounce rates. You should consult with the domain owner to determine if they recommend an alternative domain for your email use case.

Not all mail receiving organizations enforce DMARC policies. Some may use their own logic, such as quarantining messages that fail a reject policy. Some may use DMARC logic to build a domain-based reputational model based on your sending patterns even if you do not publish a policy. For example, here is a blueprint showing how you can set up custom filtering logic with SES Inbound.

If you have verified the domain identity with the legacy TXT record method, you must sign your email using a DKIM signature. The DKIM records in the DNS must be within the same domain as the domain in the From header of the messages you are signing.

If you have the domain identity verified with EasyDKIM and you also have email address identities verified within the same domain, then the email address identities will inherit the DKIM settings from the domain identity. Your email will be authenticated with DKIM and will not be subjected to the domain’s DMARC policy.

Can I use SPF instead of DKIM to align to the domain’s DMARC policy?

Messages can also pass a DMARC policy using SPF in addition to DKIM. This is enabled through the use of a custom MAIL FROM domain. The custom MAIL FROM domain needs to be a subdomain of the SES identity and the SES domain identity’s DMARC policy must not be set to strict domain alignment due to the way SES handles feedback forwarding. The domain owner enables a custom MAIL FROM domain by publishing records in the DNS. There is no way to authenticate email without publishing records in the DNS. Read Choosing a MAIL FROM domain to learn more.

The recommended approach is to use EasyDKIM primarily, and optionally enable a custom MAIL FROM domain as an additive form of authentication.

What should I do if I am not the domain owner?

The process of enabling DKIM and SPF authentication involves publishing DNS records within the domain. Only the domain owner may modify DNS for their domain. If you are not the domain owner, here are some alternative solutions.

Option 1: Segregate your email sending programs into subdomains.

This option is best for people within large or complex organizations, or vendors who are contracted to send email on behalf of an organization.

Ask the domain owner to delegate a subdomain for your use case (e.g. marketing.domain.example). Many domain owners are willing to delegate use of a subdomain because allowing for multiple use cases on a single domain becomes a very difficult management and governance challenge.

Through the use of subdomains they can segregate your email sending program from the email sent by normal mailbox users and other email sending programs. This also gives mail receiving organizations the ability to create a reputational model that is specific to your sending patterns, which means that you do not need to inherit any negative reputation incurred by others.

Option 2: Use a domain in which you are the domain owner.

This option is best if you have end-customers (or tenants) who have email addresses within domains which have domain owners that will not allow any form of delegation to you.

Use your own domain as the domain identity, and use subdomains within your domain to distinguish your end-customers from each other (e.g. tenant1.yourdomain.example, tenant2.yourdomain.example, tenant3.yourdomain.example, …). Amazon WorkMail uses this strategy for the awsapps.com domain.

This gives you complete control over the domain as well as your reputation. Use subdomains to segregate reputation between your end-customers if you have a multi-tenant business model.

Here are some additional suggestions to make your email more personable while remaining aligned to the domains’ DMARC policies.

• You may format the From header of your outgoing messages so that the display name clearly reflects the name of the message author.

From: “John Doe via My App” <john.doe_no-reply@tenant1.mydomain.example>

• Set the Reply-to header of your outbound messages so that when recipients reply, the return messages will go to the intended recipient.

Reply-to: john.doe@domain1.example

What should I do if the domain is already being used for a different email sending program?

From a deliverability perspective, it is beneficial to compartmentalize your sending into different domains, or subdomains, for different email sending programs. That will limit the reputational blast radius if something were to go wrong with one campaign. Consider using different subdomains for each sending program. For example:

• marketing.domain.example
• receipts.domain.example

DMARC was designed for marketing and transactional email use cases, so it is good practice to publish ‘reject’ DMARC policies for those subdomains. Having a strong policy doesn’t give a free pass into recipient inboxes, but it allows the mail receiving organization to know what to do with messages that aren’t authenticated, which can lead to better trust. Building trust is the best way to gain a positive reputation.

If the domain is used by normal users for day-to-day correspondences, the domain owner should be very careful about publishing a DMARC policy because it is known to create interoperability issues with mailing lists and other email providers. Many of these email domains may never publish a ‘reject’ DMARC policy. For new email sending programs, you should strongly consider using a subdomain rather than any domain that is being used for user correspondences.

Conclusion

To ensure optimal deliverability with Amazon SES, it’s essential to be aware of the domain owner’s preferences and use a domain identity for outbound messages. Keep in mind that email address identities should only be used for testing purposes or with domains without DMARC policies. Domain owners can use subdomains to segregate email sending programs, making management and governance easier while allowing mail receiving organizations to build isolated reputational models.

By following the recommendations in this blog, you’ll be better prepared to align with the domain owner’s preferences, achieve higher deliverability rates for your authenticated outbound email, and be compatible with future DMARC developments.