schedule 8-min read

DMARC Aggregate Reports Explained

DMARC AI editorial team · Last updated

How to read DMARC aggregate (RUA) reports under RFC 9990: what each XML field means, the four patterns that matter and how MSPs collect them at scale.

01

Introduction

To read a DMARC aggregate report, open the XML, note which receiver sent it and for which day, then go row by row: each source_ip with its count shows how much mail claimed your domain, and policy_evaluated shows whether it passed DMARC through aligned SPF or aligned DKIM. Every row should map to a sender you can name.

Aggregate reports, usually called RUA reports after the rua= tag that requests them, are the summaries mailbox providers send to the address in your DMARC record. Their format is now defined by RFC 9990, published in May 2026 alongside the core standard RFC 9989. This guide walks through a current report field by field and shows how to turn it into decisions. For the wider picture of the protocol, start with the DMARC guide.

02

Key takeaways

  • An aggregate report is one receiver's view of mail claiming your domain, typically for one UTC day.
  • Read it per row: source_ip, count, then the evaluated dkim and spf results.
  • DMARC passes when either SPF or DKIM passes and aligns with the visible From domain.
  • Under RFC 9990, pct is gone from policy_published and a DKIM selector is now required in auth_results.
  • MSPs that collect reports for client domains at their own mailbox need an external destination verification record.
03

What is a DMARC aggregate report?

A DMARC aggregate report is an XML file a receiving mail system sends to the address in your rua= tag, summarising every message it saw that claimed your domain in the From header. It contains counts and authentication results per sending IP, never message content. RFC 9990 describes it as receiver-originated feedback sent daily or more often.

Aggregate reports are the only DMARC data you receive continuously. Failure reports are per-message and many large receivers do not send them. Every operational choice in a rollout (fix this sender, ignore this IP, move to quarantine, hold) comes from aggregate data.

04

What does an aggregate report look like?

A report has a fixed top-level order under RFC 9990 section 3.1.1: version, report_metadata, policy_published, an optional extension, and one or more record elements. Here is a trimmed example for a fictional domain:

<feedback>
  <version>1.0</version>
  <report_metadata>
    <org_name>receiver.example</org_name>
    <email>dmarc-reports@receiver.example</email>
    <report_id>2026-11-03.example.com</report_id>
    <date_range>
      <begin>1793664000</begin>
      <end>1793750400</end>
    </date_range>
  </report_metadata>
  <policy_published>
    <domain>example.com</domain>
    <p>quarantine</p>
    <sp>quarantine</sp>
    <np>reject</np>
    <adkim>r</adkim>
    <aspf>r</aspf>
    <testing>n</testing>
  </policy_published>
  <record>
    <row>
      <source_ip>198.51.100.42</source_ip>
      <count>247</count>
      <policy_evaluated>
        <disposition>none</disposition>
        <dkim>pass</dkim>
        <spf>pass</spf>
      </policy_evaluated>
    </row>
    <identifiers>
      <header_from>example.com</header_from>
      <envelope_from>bounces.example.com</envelope_from>
    </identifiers>
    <auth_results>
      <dkim>
        <domain>example.com</domain>
        <selector>s2026</selector>
        <result>pass</result>
      </dkim>
      <spf>
        <domain>bounces.example.com</domain>
        <result>pass</result>
      </spf>
    </auth_results>
  </record>
</feedback>

The date_range values are Unix timestamps. RFC 9990 section 3.1.1.4 says a reporting period is typically a single UTC day starting at 00:00 UTC, and periods should not overlap.

05

Which fields matter most?

Six fields drive almost every decision. Read them in this order for each record:

Field Where What it tells you
source_ip row The IP that delivered the mail; map it to a sender
count row Messages from that IP the evaluation applied to
disposition policy_evaluated What the receiver did: none, quarantine or reject
dkim policy_evaluated DKIM result after the alignment check
spf policy_evaluated SPF result after the alignment check
header_from identifiers The visible From domain the report is about

The auth_results block holds the raw DKIM and SPF results "uninterpreted with respect to DMARC" (RFC 9990 section 3.1.1.11). Compare it with policy_evaluated when debugging: a raw DKIM pass for vendor-mail.example next to an evaluated dkim fail means the signature is valid but signed with the wrong domain, so it does not align.

06

What changed in reports with RFC 9990?

Three changes matter for anyone parsing reports. First, policy_published no longer carries pct, because RFC 9989 removed that tag; it can now carry np (the policy for non-existent subdomains), testing (the value of the t tag) and discovery_method. Second, a DKIM selector is now required whenever a signature is reported. Third, reports may carry extension elements, so parsers must tolerate elements they do not recognise.

If your parser was written for RFC 7489 reports, test it against a current report before trusting its numbers. RFC 9990 explained for MSPs covers the standard in more depth.

07

How do you interpret the results?

Four patterns explain most rows. DMARC passes when at least one of SPF or DKIM passes with alignment, so the first two rows below are passing mail.

Evaluated SPF Evaluated DKIM Usual meaning Action
pass pass Legitimate, fully authenticated sender None
pass fail Sender signs with its own domain, or not at all Set up aligned DKIM: it survives forwarding, SPF does not
fail pass Usually forwarded mail where DKIM survived Normally safe to leave
fail fail Spoofing, or a legitimate sender not yet configured Identify the IP before enforcement

The work of a rollout is moving every legitimate source_ip into the first row. When every named sender passes consistently, you have the evidence to move from monitoring to enforcement.

08

How are reports delivered?

Receivers email the report as an attachment. RFC 9990 section 3.5.2 says the XML should be compressed with GZIP and gives a filename pattern of receiver, policy domain, begin and end timestamps, for example mail.receiver.example!example.com!1013662812!1013749130.xml.gz. A re-sent report reuses its original filename, which is how a parser can deduplicate.

How do MSPs collect reports for many client domains?

When a client domain sends its reports to an address on the MSP's domain, receivers first check that the MSP's domain agrees to receive them. Under RFC 9990 section 4 the MSP publishes a TXT record such as clientdomain.com._report._dmarc.msp-reports.example containing v=DMARC1. A wildcard record at *._report._dmarc.msp-reports.example authorises reports for any domain. Without it, many receivers silently drop the reports, which looks exactly like "no reports arriving".

09

Best practices

  • Name every sender. Map each source_ip to a service: mail platform, CRM, billing system or attacker. If you cannot name it, that is the work.
  • Do not dismiss low-volume rows. A row with count of 1 from a new IP can be a probe or a newly added tool.
  • Track pass rate per sender over time. A sender at 100% one week and 95% the next is drifting.
  • Expect gaps. Not every receiver reports, and coverage varies by day; see why DMARC reports are not always perfect.
  • Parse with software past a handful of senders. Raw XML is fine for spot checks; for daily operations use a parser or platform. How to read DMARC XML reports covers manual reading, and whether AI can analyse DMARC reports safely covers what a chatbot can and cannot do with them.
10

Pick yesterday's largest report, list every source_ip, and name each one. Any IP you cannot name is either a sender to authenticate or abuse to watch. If no reports arrive at all, check the rua= address, and if it sits on another domain, check the _report._dmarc authorisation record described above.

11

FAQ

How often do receivers send aggregate reports?

Typically once a day. RFC 9990 describes daily or more frequent reports covering a single UTC day. The ri interval tag from RFC 7489 is not part of RFC 9989, so you cannot request a different interval.

Why don't I see reports from some receivers?

Not every receiving system sends DMARC reports. If a large provider is missing, check that your rua= mailbox accepts the mail and, for an address on another domain, that the external destination verification record exists.

Do aggregate reports include message content?

No. They contain IP addresses, counts, domains and authentication results only, never subjects or bodies. Per-message detail belongs to failure reports, defined separately in RFC 9991.

What is the difference between auth_results and policy_evaluated?

auth_results holds the raw SPF and DKIM outcomes for the domains checked. policy_evaluated holds the DMARC view: whether each mechanism passed with alignment to the From domain, and what the receiver did with the mail.

Can I still rely on reports written for RFC 7489?

Older reports remain readable, but fields have changed: pct has gone and selectors are required for DKIM results. Check that your parser handles both while receivers move to the new format.

12

Final thoughts

Aggregate reports are the operational core of DMARC. Read them row by row, name every sender, and the rollout becomes a sequence of evidence-based decisions instead of guesses. For an MSP, the extra step is infrastructure: an authorised reporting domain and a parser that copes with RFC 9990 reports from many receivers at once.

Related articles

Related tools

Ready to Implement?

Get authenticated mail moving in minutes: start free, book a guided demo, or talk to the team about your stack.