To find out who sends email from your domain and who owns each stream, you need an inventory that maps every sender to a real service, owner, and business purpose. Tie each sending service to an approver who authorizes that traffic, to an operator who configures DNS and mail parameters, and to whatever message evidence you have. Where an owner or an authentication result remains undocumented, record that gap explicitly instead of leaving it out; the next engineer reviewing the register needs to see what remains uninvestigated.
Start with an agreed list of visible From domains and subdomains, access to application administrators and mail operations teams, and current configuration records, diagnostic messages, and aggregate reports. The method below is a working practice I recommend; its register fields are not a schema any RFC requires.
Identifying services that should send from your domain
Interview application owners and postmaster teams to catalog every account, relay, cron job, and SaaS tool delivering mail on behalf of each domain. Security alerts, transactional receipts and marketing broadcasts all belong on the list when they fall within scope. For every lead, record who identified the service and where to inspect its active configuration. Existing DNS records and vendor dashboards are a good starting point, but they do not establish who the current business owner is.
Create a distinct record for each service instance and business purpose. Split entries whenever administrative ownership, technical operators, or authentication configurations diverge. This administrative work is not a formality: the operational guidance in RFC 7208 points out that identifying permitted relays requires understanding the sending architecture, not only the technical configuration.
For every inventory entry, capture:
- The service name, account identifier, business function, sending cadence, and intended
Fromdomain. - The business approver, meaning the decision-maker who can justify why the email must send and describe the impact if delivery halts.
- The technical operator responsible for application configuration, alongside the DNS administrator if administrative boundaries are split.
- Direct references to approval tickets and technical documentation, writing down an explicit unknown wherever proof is missing.
A generic provider brand name cannot replace a specific tenant or account identifier. Before treating an association as established, ask the identified operator to confirm the exact instance and its purpose.
What DMARC aggregate reports reveal (and what they hide)
When analyzing an observation from an aggregate report, record the reporting organization and the covered reporting window alongside the connecting IP, message volume, evaluated From domain, and authentication findings. A connecting IP address alone cannot identify an individual application account. Some report fields are optional under the standard, so missing data points should remain marked as unknown rather than guessed (RFC 9990, report records). This analysis assumes reports reach an active ingestion inbox; if processing is not yet operational, start with processing DMARC aggregate reports at scale.
Evaluate every authentication verdict alongside the specific identifier evaluated by receiving servers. Under RFC 7208, an SPF pass attaches strictly to an SMTP envelope identity, so record that identity rather than substituting the visible From domain. For DKIM, document the signing domain from d=, the selector from s=, and the signature validation state. The presence of a signature header does not mean verification passed, nor does it identify who authorized the stream (RFC 6376, signature fields).
Keep DMARC alignment records distinct from underlying SPF and DKIM checks, as these fields address separate verification layers (RFC 9990, authentication results). Under an applicable policy, DMARC requires a passing SPF or DKIM identifier to align with the visible From domain; an exact domain match satisfies strict alignment requirements (RFC 9989, identifier alignment). For SPF, DMARC looks only at the SMTP MAIL FROM identity (RFC 9989, SPF identity for DMARC).
For evidence taken from a received message, trust Authentication-Results only once you have established who produced the header and how it reached you. The authserv-id token identifies the validating host, but encountering a familiar hostname in header text proves nothing by itself, because untrusted intermediaries can forge incoming headers (RFC 8601, trust boundaries).
Converting unverified streams into diagnostic requests
Treat unfamiliar report entries as investigative leads until you can corroborate both platform ownership and legitimate business purpose. Track down the specific account ID and find the engineer or admin who manages it. You can’t rely on aggregate reports alone to prove ownership, since report data itself can be spoofed (RFC 9990, forged reports).
When a recognized service has generated no diagnostic samples, ask its technical operator to trigger a controlled test message to an authorized diagnostic mailbox. Establish a safe trigger beforehand to avoid executing accidental transactions or emailing real customers. Save the raw message headers and reference documentation, then inspect the resulting authentication verdicts. If the delivered message cannot be traced back to the service account, or if the provenance of the header remains questionable, document what stays unresolved. This is a proposed diagnostic procedure; no test was performed for this article.
The synthetic example below features fictional team members, services, and evidence. Assumed outcomes rely on valid DMARC policies and non-null SPF MAIL FROM identities. It demonstrates three distinct follow-up paths rather than an exhaustive enterprise inventory.
| Service and purpose | Approval and responsibility | Assumed message evidence | Next evidence request |
|---|---|---|---|
orders-prod: purchase receipts | Rin approves the use; Dev operates the service | From: receipts.example.com; SPF passes for mailer.example.net; DKIM passes with d=receipts.example.com, s=receipt-a | Retain the sample reference and identify other untested sending paths. DKIM supplies an exact aligned pass here. |
ops-east: failure alerts | Business approver unknown | From: alerts.example.com; DKIM passes with d=alerts.example.com, s=alert-a | Obtain the relay account record and an accountable approver. A passing sample leaves ownership unresolved. |
renewal-batch: annual service notices | Sol confirms the job is required | Expected From: notices.example.com; no sample or matching report observation | Ask the operator for a safe representative test and the timing of the next scheduled event. Authentication evidence remains unknown. |
In the purchase receipts row, the passing SPF identity does not align with From; aligned DKIM supplies the passing DMARC verdict. In the alerts row, a passing cryptographic signature cannot replace a documented business approval. That is why business authorization and message evidence remain separate inventory fields.
Packaging the register with its open questions
Group open questions by service, owner, authentication status, and monitoring coverage, and assign each unresolved item to an engineer with clear next steps. Unobserved sending jobs, especially infrequent annual or quarterly runs, belong on this tracking list.
A lack of report data doesn’t mean a service has stopped sending email; mailbox providers aren’t required to send aggregate reports (RFC 9989, reporting participation). Evaluation timelines depend on sending rhythm: a single test or a fixed monitoring window cannot confirm that every outbound route has been audited (RFC 9989, assessment before enforcement).
An authentication failure may involve legitimate mail. Figure out what business function the email serves and work with the service owner to fix the authentication issue before turning on enforcement (RFC 9989, remediation). The register then supports the separate DMARC enforcement decision.
Track findings directly in the sender register worksheet. For deeper report analysis, continue with interpreting a DMARC aggregate report.
Technical sources were checked on 8 September 2026 and include the May 2026 DMARC standards. Their publication does not establish how any particular receiver or reporting product implements them.
Frequently asked questions
Can a connecting IP in a DMARC report identify the sending application?
No. A connecting IP on its own does not identify an application account, and because some report fields are optional, an absent value stays unknown rather than becoming a guess.
Does a passing DKIM signature confirm who approved the sending service?
No. The presence of a signature header does not establish a pass, and even a passing DKIM result cannot supply the missing business decision; it does not identify a business approver.
Can you trust the Authentication-Results header on an incoming message?
Only after verifying the authenticity of the reporting mail server and the network path it traversed. Its authserv-id names the checking service, but a familiar hostname in the header text is not enough on its own, because the field can be forged.
Does an absent report observation mean a sender has stopped transmitting?
No. Mailbox providers are not obligated under RFC 9989 to return aggregate reports for every processed message stream.
Which identity belongs in your SPF inventory records?
The SMTP identity that the SPF pass belongs to, not the visible From domain substituted for it. On the SPF side, DMARC uses only the MAIL FROM identity.


