How Do You Read DMARC Aggregate Reports and Spot Spoofing?
Published 2026-11-14 · Reviewed 2026-11-14 · 6 min read · By the Ingenious AI team
Is someone impersonating your domain right now?
Run the free DomainGuard scan: lookalike registrations, MX-armed variants, certificate-transparency and CAA checks for your domain, in one report.
Run my free DomainGuard scanThe short answer
You read a DMARC aggregate report by answering three questions per sending source: who is this, is their mail aligned, and what did receivers do with the failures? The reports arrive daily as XML from major mailbox providers, listing every IP that sent mail using your domain, with counts and SPF and DKIM results. Humans are not meant to read raw XML at scale — use a report parser or service — but the logic is simple: known sources failing alignment are configuration work; unknown sources failing everything are spoofing, and at p=reject they are spoofing that failed.
What actually arrives in your mailbox
Each report covers roughly a day from one receiver — Google, Microsoft, Yahoo, and others each send their own — as a compressed XML attachment. Inside, a header names the reporting organisation and window, a policy block repeats the DMARC record the receiver evaluated against, and then a series of records: one per combination of sending IP, disposition, and authentication results, each with a message count. One busy domain can generate thousands of lines that summarise into a handful of sources.
Two implications follow. First, coverage is partial by design: you see reports from receivers who choose to send them, covering a large share of consumer mailbox traffic but never literally every message. Second, the data is sampled reality, not a log of your mail server — forwarded mail, for instance, appears with the forwarder's IP. Read reports for patterns and sources, not as an accounting ledger, and they will never mislead you.
The three questions, source by source
Question one: who is this source? Map each sending IP and authentication domain back to a service — your mailbox provider, your CRM, the invoicing tool — using reverse DNS, the DKIM signing domain, and simple searching. Maintain the mapping in a sender register; within a month, ninety percent of your report volume will be familiar names, and the unfamiliar remainder is where your attention belongs.
Question two: is it aligned? For each known source, at least one of aligned SPF or aligned DKIM should pass consistently. A source passing neither is a fix task: branded DKIM configuration or return-path alignment, verified in the following days' reports. Question three: what happened to failures? The disposition field shows whether receivers applied your policy — none, quarantine, reject — which is how you confirm enforcement is real in the world and not just written in your DNS.
Spotting actual spoofing in the noise
Spoofing has a signature shape in aggregate data: sources you have never seen, sending from infrastructure unrelated to any vendor, failing both SPF and DKIM alignment, often in bursts and often claiming your exact domain. Before enforcement, those rows show disposition none — the messages were delivered, and your reports are the only witness. After p=reject, the same rows show disposition reject, which is the sound of the door closing; the attempts rarely stop, but they stop arriving.
Distinguish spoofing from your own misconfigurations by the alignment pattern. Your broken vendor fails alignment but sends recognisable, steady volume with your real DKIM selector or your known return-path. Spoofers fail everything, sign as nothing you own, and cluster in short campaigns. Keep a simple log of spoofing sources and volumes month to month; a sudden order-of-magnitude spike is worth a look for a live campaign and a check that no lookalike infrastructure is involved on the side.
Tools: nobody should read raw XML weekly
Parser services — including the reporting built into DomainGuard's monitoring — turn the XML stream into the three-question view: sources ranked by volume, alignment state per source, failures dispositioned, new sources flagged. Free tiers exist for small domains and are entirely adequate; the expensive mistake is the unmonitored rua mailbox filling with attachments nobody opens, which is where most domains' reports actually live.
Set the rua address deliberately: a dedicated mailbox or the reporting service's provided address, never a person's inbox that will be abandoned at the next role change. Reports are a data feed with an owner, like any other telemetry. Give them a named reviewer and a monthly slot, and the feed will repay you the first time it names a sender your IT team forgot existed — which, in practice, happens in the first month for almost everyone.
From reports to decisions
The monthly review should end with a short list of actions, not a sigh at a dashboard. Typical outputs: a new vendor to align before its contract renews; a departed tool still sending, to be cut off; a subdomain that started sending without records, to be brought into the fold; spoofing volume noted and enforcement confirmed holding. Twice a year, the review should also ask whether subdomains and parked domains all still carry the right records, because estates drift at the edges first.
Treat the reports, finally, as your proof. When a customer or an enterprise prospect asks how you know your email cannot be spoofed, the aggregate data is the answer: these are the only sources that send as us, this is their alignment, and this is what receivers did with the forgeries last quarter. Few small companies can produce that evidence. Having it puts you in the minority that can — on security questionnaires and in incident rooms alike.
Five actions to take today
- Point your DMARC rua tag at a parser service or dedicated mailbox, and name a monthly reviewer.
- Build a sender register mapping every report source to a service and an internal owner.
- Fix any known source passing neither aligned SPF nor aligned DKIM, verifying in later reports.
- Classify unknown all-failing sources as spoofing; confirm dispositions show quarantine or reject once enforced.
- End each monthly review with a written action list — new senders, departed tools, uncovered subdomains.
Worked example: Harbor Lane Accounting finds a ghost sender
Harbor Lane Accounting, a fictional firm monitoring 6 domains, began monthly aggregate reviews after enforcing DMARC. Month one surfaced a surprise: a document-scanner service still sending statements as the firm from two offices, unaligned — a tool IT believed had been retired the previous year. Month two showed 340 spoofed messages from three unknown sources, all dispositioned reject under the firm's p=reject policy. The ghost sender was properly retired, the spoofing required no action beyond being logged, and the firm's quarterly summary to its insurer contained something few firms its size can show: per-source proof of who sends as them, and what happened to everyone else who tried.
FAQ
Why are the reports in XML instead of a readable format?
They are designed for machine processing at receiver scale. Every DMARC reporting tool exists to translate them; pick one and never read raw XML except to satisfy curiosity once.
Do aggregate reports show message content or recipients?
No. They contain sending sources, counts, authentication results, and dispositions — metadata only. That privacy-preserving design is why receivers are willing to send them at all.
How is this different from forensic (ruf) reports?
Forensic reports are per-message failure samples and are rarely sent anymore for privacy reasons. Aggregate reports are the daily statistical feed everyone actually uses for operations.
A source I do not recognise passes alignment. Should I worry?
Yes — investigate promptly. Unknown sources that pass usually mean a forgotten legitimate tool or an over-broad SPF record authorising infrastructure you no longer control. Both need resolving; the register will tell you which.
How long should I keep report history?
At least a year. Trends — spoofing volume, new sources, seasonal senders like payroll year-end tools — only become visible across quarters, and incident investigations routinely ask who was sending as you months ago.