[ Standards & DMARCbis ]

DMARCbis readiness: what to review before changing DNS

DMARCbis keeps v=DMARC1. A useful readiness review checks changed policy behaviour, receiver support, reporting and the legitimate mail your organisation depends on.

By the EmailAuth research team · 9 min read

[ Key points ]

  • The record version remains v=DMARC1.
  • Testing, domain-boundary handling and reporting need explicit expected-result checks.
  • A published standard and a deployed implementation are different things.
On this page

DMARCbis changes how parts of DMARC are specified and implemented. It does not ask you to rename your record to DMARC2. The core specification, RFC 9989, was published in May 2026, while RFC 9990 and RFC 9991 cover aggregate and failure reporting separately.

The practical task is to work out which changes affect your domain, your sending services and the software that processes your authentication data. That starts with understanding the current specifications, then comparing them with the configuration and processes already in use. The result should be a clear readiness register with evidence, owners and a decision for each finding.

Map the three specifications to your systems

Use the current documents for their respective jobs:

DocumentWhat to review
RFC 9989Core DMARC evaluation, policy discovery, alignment and record semantics
RFC 9990Aggregate reporting format, generation and processing
RFC 9991Failure reporting and its handling considerations

Start with the existing record and configuration history. Do not delete an operating record simply because an older guide cites RFC 7489. Identify the instructions that changed and the software that implements them.

The main review areas include the following:

  • Domain discovery: RFC 9989 section 4.10 defines DNS tree walking. Review systems that assume an older public-suffix-list discovery model.
  • Policy tags: Check the current IANA registry. The pct, rf and ri tags are historic. Assess t, np and psd by their actual scope rather than adding them as an upgrade checklist.
  • Testing: Establish how the relevant evaluator implements the binary t signal. A declared testing state and a receiver's final disposition are different observations.
  • Reporting: Confirm that collectors and parsers handle the report formats and fields you expect to receive, without losing relevant legacy observations.

The version marker remains v=DMARC1. Do not substitute DMARC2 or mechanically replace an old percentage with a testing flag.

Translate the changed tags into tests

The t flag defaults to n, requesting application of the declared assessment to DMARC failures. Setting t=y signals that the owner is testing rather than asking for the declared assessment to be applied directly. RFC 9989 section 4.7 defines a one-level reduction for failing messages: reject becomes quarantine, and quarantine becomes none. Implementations may also have relevant special handling.

The flag does not change a failed authentication result into a pass, alter report generation, or affect a none policy. It is binary, not a request to sample an arbitrary percentage of mail. Establish implementation support before relying on it for real traffic, and keep receiver-local filtering separate from the policy-level test result.

Two other tags affect the scope of the assessment:

  • np: Specifies the assessment for non-existent subdomains of the prevailing organisational domain. If absent, the applicable fallback is sp when present, otherwise p. A missing MX record alone does not establish non-existence.
  • psd: Marks domain-boundary information used during discovery. The default u leaves boundary determination to the discovery rules; y identifies a public suffix, while n identifies a non-public-suffix organisational boundary for itself and its subdomains. Do not set either explicit boundary value without establishing that it describes the domain correctly.

An ordinary organisation should not label its domain as a public suffix merely to claim readiness. Review actual discovery assumptions and delegated namespaces. The subdomain policy guide covers the narrower inheritance question.

Identify the mail that must keep working

A domain may represent much more than employee email. It can appear on invoices, login links, customer alerts and messages sent by outside agencies. A standards update is an opportunity to check those dependencies before a rushed configuration change disrupts them.

Ask for three deliverables: an inventory of the services using your identity, an explanation of the specific changes that affect you, and a test plan for critical messages. The answer should distinguish a published specification from a feature already deployed by a particular receiver or software service.

Allocate ownership as well as budget. A report-processing change needs someone responsible for the reporting pipeline. A sender that cannot authenticate correctly needs its business owner involved. Calling both items a DNS upgrade hides the work.

Build a readiness register

The following is an illustrative planning register, not an assessment of a real domain:

FindingWork requiredEvidence of completion
Rollout instructions assume percentage samplingTest the binary flag using the controlled fixtures belowRecorded declared assessment, assessment after the testing flag and independent receiver decision
Report parser assumes one legacy schemaReview accepted formats, fields and error handlingVersioned valid, legacy and malformed fixtures
Critical application has no known ownerEstablish the business service and supported identitiesNamed owner and verified received-message sample
DNS editor is controlled by an outsourced teamAgree approval and recovery responsibilitiesExact change preview, approval and restoration material

The following synthetic fixtures specify expected results for a controlled evaluator implementing RFC 9989. They are test inputs, not production record templates or tests already performed. Assume a valid applicable policy is discovered and the test message fails DMARC.

FixtureInputExpected policy-level observation
Testing flagp=reject; t=yValidation remains fail; assessment after t=y is quarantine, rather than the declared reject
Flag at nonep=none; t=yThe flag leaves the none assessment and report-generation behaviour unchanged
Non-existent childA prevailing organisational-domain policy has p=none; sp=quarantine; np=reject, with default t=n; the child is established as non-existent and has no direct policyThe applicable child assessment is reject, rather than the existing-child sp value

Capture the evaluator version, input, observed validation, selected assessment and any final handling separately. A controlled fixture establishes the behaviour tested; it does not prove every Internet receiver implements the same rules. A policy-level quarantine observation is not a universal promise of spam-folder placement.

Review policy suitability as well. RFC 9989 section 7.4 advises against reject for general-purpose email domains because legitimate indirect routes, including mailing lists, can break. A readiness review should identify those routes and assess the applicable interoperability guidance before proposing stronger policy.

Keep configuration changes separate from software or process changes. Some findings need no immediate DNS edit. Others require a provider update, a parser fix or a controlled implementation test.

Check interoperability and the limits of a readiness claim

A readiness claim should name its scope. Does it cover the domain portfolio, the policy evaluator, the report processor, or only a record check? These are different assessments.

Use the register to track implementation versions and unresolved differences. Select applicable cases from existing and non-existent subdomains, alignment modes, indirect mail and DNS failures.

Treat receiver discretion as part of the operating model. Publication of a specification does not make every receiver behave identically. A published reject preference is also not evidence that every recipient rejected every failing message.

Maintain independent controls for compromised accounts, deceptive display names, lookalike domains and sensitive business approvals. DMARC readiness does not turn domain authentication into a verdict on the safety of a message's content.

Make and verify the required changes

Before a production change, preserve the previous state, identify affected flows, obtain the appropriate approval and define recovery conditions. Then verify both the published configuration and representative messages through the complete sending route.

If a result disagrees with the expected behaviour, pause expansion to other domains and resolve or explicitly accept the compatibility finding.

For the underlying identities, use the alignment diagnosis guide. For an initial configuration, follow the DMARC creation and verification workflow. For the reporting layer, use the aggregate-report walkthrough.

The practical takeaway

Close the assessment when each applicable finding has an owner, a supported decision and reproducible evidence. Keep unresolved implementation differences visible, verify critical messages and report processing, and retain the approved recovery conditions. A readiness label should point to that record, not replace it.

[ By role ]

What this means for you

Choose your role for practical next steps.

IT / Mail Admin

Build a reproducible gap register.

Compare the current core and reporting specifications with the actual evaluator, records and parsers in use. Separate a documented standards change from evidence that a relevant receiver or software version implements it.

Next steps

  1. Capture the existing configuration and software versions.
  2. Test applicable discovery, alignment and reporting cases.
  3. Keep the proposed change, test result and recovery route together.

CISO / SecOps

Approve a defined scope, not a readiness badge.

A record check, domain-portfolio review and parser assessment are different scopes. Require the assessment to state which it covers, where evidence is missing and how interoperability and residual risks are handled.

Next steps

  1. Review indirect-mail and general-purpose policy implications.
  2. Ask for implementation and compatibility evidence.
  3. Retain account, lookalike-domain and business-approval controls.

Founder / Non-technical

Ask what the update changes for the business.

Request a scoped plan for the communications customers and staff rely on. A standards update may require work in a sender, reporting system or operating procedure as well as DNS. Fund and assign those specific tasks.

Next steps

  1. Identify the owner of billing, login and notification mail.
  2. Ask which findings affect your services and why.
  3. Require a verification result before accepting an upgraded status.

[ FAQs ]

Frequently asked questions

Do I need a DMARC2 record to use DMARCbis?

No. RFC 9989 retains v=DMARC1. Review the changed protocol, policy and reporting behaviour rather than replacing the version with DMARC2.

Is DMARCbis still only a draft?

The core revision was published as RFC 9989 in May 2026, with RFC 9990 and RFC 9991 addressing reporting. Check the current document status and implementation support for the feature you plan to use.

Should I remove pct from every record immediately?

Review the current registry, the intended rollout behaviour and relevant implementations first. The tag is historic under the updated specification, but a mechanical edit is not a substitute for a tested transition.

Is the t tag a replacement for percentage-based rollout?

The t flag is binary and defaults to n. With t=y, RFC 9989 specifies a one-level reduction for failing messages: reject to quarantine, or quarantine to none. It does not change a DMARC failure into a pass, affect a none policy or alter report generation. Confirm implementation support and keep final receiver handling separate.

Does an RFC update mean my reports need a parser update?

It means the parser should be assessed. Test the formats, fields, extensions and legacy reports it must handle before deciding whether an implementation change is necessary.

What should a small business ask its provider to deliver?

Ask for a scoped gap register, a list of affected senders and systems, the proposed changes, verification evidence and a recovery route. Avoid accepting an unexplained upgraded status as the deliverable.

Can a DNS checker certify DMARCbis readiness?

A DNS check can assess particular published configuration. It cannot by itself validate a report parser, every sending flow, receiver implementation or the organisation's change process.

What evidence should close a DMARCbis readiness review?

Keep the applicable findings, owners, approved changes, evaluator or parser versions, expected and observed results, critical-message verification and recovery conditions. Record unresolved compatibility differences rather than closing them with an unexplained readiness label.

[ Sources ]

Sources

  1. RFC 9989: DMARC core protocol
  2. RFC 9990: DMARC aggregate reporting
  3. RFC 9991: DMARC failure reporting
  4. IANA DMARC parameters registry

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