[ DMARC & policy ]

Reassess DMARC when you add an application or email service

Onboard each sending workflow with an owner, supported identities, a policy decision and a received-message test before relying on it for business mail.

By the EmailAuth research team

[ Key points ]

  • Approve each meaningful workflow, not just a provider console.
  • Use supported identity settings and explain the exact DNS diff.
  • Keep untested paths pending and plan recovery and retirement.
On this page

Adding an email service changes your sending estate even when the main DMARC record stays untouched. A newly purchased billing platform, support desk or marketing application can display your domain in From while authenticating a different identity. Treat onboarding as a change to a business mail route, not just a request to paste DNS records.

The objective is to establish an owned, supported and tested sending path before customers depend on it. The review can be small, but it should cover every workflow the application will use and the evidence needed to maintain it later.

Define the intended sending scope

Ask the application owner which messages the service will send, which From addresses users will see, who receives them and whether the application has separate transactional and marketing paths. Include low-frequency operations such as account recovery and scheduled statements.

Record who controls the provider account and who can change its sender settings. Keep account security and business approval alongside technical authentication. DMARC does not establish that a particular campaign, invoice or request was authorised by your organisation.

Choose the intended domain deliberately. A subdomain can provide an operational boundary, but its policy and alignment behaviour must be checked. Do not assume the main domain's enforcement policy automatically gives every child workflow the same outcome. Use the subdomain policy guide when planning that boundary.

Obtain the provider's actual configuration

Use current first-party instructions for the specific product and sending mode. Ask what will appear in From, which envelope domain the service uses, and whether it provides customer-domain DKIM signing. Domain ownership verification alone does not establish that messages are signed with an aligned identity.

Copy generated DNS values from the authorised service configuration rather than from an example article. Check the exact DNS owner names, record types, existing records and zone ownership before making a change. A selector or verification token from another account is not a reusable template.

If the service needs an SPF change, establish which domain's SPF is evaluated and how the change affects its lookup budget under RFC 7208 §4.6.4. Under RFC 7208 §2.4 and RFC 7208 §4.4, adding the vendor to the organisation's root SPF may be irrelevant when the service uses a different envelope domain. Keep the requested change and its rationale together.

Make the readiness matrix specific

The following is a synthetic planning example, not observed test data. Its “pending” cells are deliberate: configuration claims must be replaced with evidence before approval.

  • Workflow: Receipts; Intended From domain: billing.example.com; Evidence needed: Received receipt with aligned DKIM, envelope identity and policy discovery recorded; Release decision: Pending test
  • Workflow: Password reset; Intended From domain: accounts.example.com; Evidence needed: Reset message through the real application path, including any gateway; Release decision: Pending test
  • Workflow: Campaigns; Intended From domain: news.example.com; Evidence needed: Actual campaign message, authentication results and applicable recipient-provider requirements; Release decision: Pending review

Add the owner, provider account reference, source documentation, last test date and rollback action to your working register. Separate workflows whenever the service uses different sender identities or delivery infrastructure. One successful receipt does not certify a campaign path.

Review the policy diff, including no change

Capture the existing policy and explain whether onboarding changes it. “No DMARC policy change” is a legitimate outcome if the new route authenticates and aligns under the existing policy. It is still a decision worth recording.

DMARC's message-level pass condition requires a passing aligned identity. For an enforcement deployment, also consider route resilience: RFC 9989 says a domain owner must not rely solely on SPF when the Author Domain policy is p=reject. Verify supported aligned DKIM rather than relying on a direct-mail SPF test alone. See RFC 9989 §5.3.5 and RFC 9989 §8.

Do not lower a parent policy by default to accommodate an unready application. First resolve the sender's supported identity configuration. If a temporary exception is necessary, obtain a scoped approval with an owner, expiry and exit test. Record the precise risk and affected domain rather than a blanket promise that delivery will improve.

Test the route the business will use

Send non-sensitive examples to authorised recipients through each workflow. Record the trusted receiving system, From domain, SPF identity, every relevant DKIM result, applicable policy and DMARC outcome. Test after any gateway or forwarding stage that belongs to the intended route.

Check actual delivery separately. A DMARC pass does not guarantee inbox placement or satisfaction of all recipient-provider rules. For instance, Gmail's bulk-sender requirements include SPF, DKIM and DMARC configuration, as well as additional delivery requirements. Consult the current Gmail sender guidelines when that audience is relevant.

Approve only the workflows that have sufficient evidence. If a low-frequency job cannot yet be exercised, retain it as pending with a scheduled owner action rather than marking the whole application complete.

Monitor and retire deliberately

Compare subsequent aggregate reports and application delivery logs with the expected sender pattern. Reports can reveal traffic, but cannot by themselves prove the provider account was authorised or that every rare workflow was exercised.

Define recovery before launch: disable the affected sending workflow, restore the approved sender configuration or route through a previously verified service where supported. DNS rollback has caching consequences, including qualifying stale-answer use after TTL expiry under RFC 8767 §4, so it is not an instant universal reversal.

Finally, set a retirement condition. When the application is removed, review its sending permissions, provider account and DNS dependencies before cleanup. Keep the sender inventory current through purchase, change and retirement, not only on the day DMARC enforcement is enabled.

Make the next decision explicit

In the planning fixture, a verified receipt path would not clear the reset or campaign rows. Record each actual result and the evidence date, or retain “pending.” For a no-change policy decision, still record the identity configuration and tested route. The planning examples do not assert that any application account, DNS record or message test exists.

[ Takeaway ]

Approve each meaningful workflow, not just a provider console. Use supported identity settings and explain the exact DNS diff. Keep untested paths pending and plan recovery and retirement.

[ By role ]

What this means for you

Choose your role for practical next steps.

IT / Mail Admin

Approve the route, identities and DNS diff together.

Use provider-generated values and current documentation, then test the production sending path. Keep a matrix when the application has several authentication modes.

Next steps

  1. Check existing DNS and lookup constraints before edits.
  2. Capture the trusted receiver result for each workflow.
  3. Plan a scoped recovery route and post-launch checks.

CISO / SecOps

Make new sending authority an accountable change.

A new service gains a role in communications carrying the organisation's identity. Require a business owner, controlled account access and an evidence-backed exception decision where necessary.

Next steps

  1. Keep temporary exceptions narrow and time-bound.
  2. Review dependencies hidden by a single successful test.
  3. Remove obsolete authority when the service is retired.

Founder / Non-technical

Include email readiness in the software launch.

The purchase is not complete when a console turns green. Ask the owner to demonstrate the receipts, recovery messages or notifications that customers will depend on.

Next steps

  1. List every business-critical sending workflow.
  2. Hold untested workflows as pending.
  3. Assign ownership through retirement, not just launch.

[ FAQs ]

Frequently asked questions

Must we change DMARC whenever we add an email service?

No. The new route may work under the existing policy. Record that decision and verify the actual messages rather than changing policy automatically.

Does domain verification prove the new service passes DMARC?

No. It may only prove control of a domain. Test the actual From, envelope and signing identities and the receiving system's evaluation.

RFC 9989 §5.3.5

Should the service always be added to our root SPF record?

Only make the supported, justified change for the identity SPF actually evaluates. A service using another envelope domain may not be affected by your root SPF record.

RFC 7208 §2.4 RFC 7208 §4.4

Can one test certify all workflows in the application?

No. Transactional mail, campaigns and recovery messages may use different identities or routes. Test each meaningful workflow or retain it as explicitly pending.

Should we move the parent policy to none during onboarding?

Not by default. Resolve the sender configuration first. Any temporary exception needs defined scope, approval, an owner, an expiry and an exit test.

Is an SPF-only pass enough under p=reject?

It can explain a particular message's DMARC pass, but RFC 9989 says a domain owner must not rely solely on SPF with p=reject. Verify aligned DKIM and the intended routes.

RFC 9989 §5.3.5 RFC 9989 §8

What belongs in the approval record?

The workflow owner, provider configuration, policy diff or no-change decision, actual test results, outstanding limitations and recovery action.

What happens when the application is retired?

Review its account, sending permissions and DNS dependencies, then remove obsolete access and records through the approved change process. Avoid breaking shared dependencies blindly.

[ Sources ]

Sources

  1. RFC 9989

    Sections §4.4 · §5.3.5 · §5.4 · §8 · §3.2.2

  2. RFC 7208

    Sections §2.4 · §4.4 · §4.6.4

  3. RFC 8601

    Sections §7.1

  4. RFC 8767

    Sections §4

  5. Google Gmail sender guidelines

    Sections Sender requirements and guidelines: bulk senders · Email authentication requirements and guidelines

[ Glossary ]

Glossary

DMARC
A mechanism that evaluates whether a passing authenticated identifier aligns with the Author Domain under the applicable policy.

RFC 9989 §5.3.5

Author Domain
The domain in the message From header field, used as the DMARC reference identity.

RFC 9989 §3.2.2

Enterprise provenance

Why top financial institutions and governments retired their email security dashboards.

Traditional email authentication platforms sell you a dashboard and ask your engineers to do the labour. In high-stakes environments, manual oversight means delayed enforcement, configuration drift and unmonitored shadow domains.

EmailAuth is the agentic email provenance platform. Our agents enforce p=reject, keep third-party senders aligned and move your domains to full RFC 9989 compliance, the new DMARC standard,, with human-in-the-loop auditability.

Governments Large enterprises Tier-1 financial institutions

SaaS · Dedicated tenancy · Air-gapped on-prem · BYOK encryption

Knowledge Base