Tool guideEmail Tools

APRF Report Analyzer – Guide to Aggregate Performance Reports

How APRF reports (draft-ietf-mailmaint-aprf) are structured, what segments and SDI are, why counts must be read exactly as reported, and how the NetTools.tech APRF Report Analyzer works.

By NetTools.tech · · 4 min read

DMARC aggregate reports tell you whether your mail authenticates. They do not tell you where it lands: inbox or spam, and whether recipients engage with it. APRF — the Aggregate Performance Reporting Format — is a proposal for mailbox providers to report exactly that, per DKIM domain and selector, once a day.

The APRF Report Analyzer reads these JSON reports in your browser and shows every count exactly as the provider reported it, segment by segment, with a validation against the current draft.

What is APRF?

In simple terms

An APRF report is a small JSON file from a mailbox provider: for one DKIM signing domain and selector and one UTC day, how many messages went to the inbox, how many were unwanted, and how recipients engaged — optionally split into segments you define.

Technical background

APRF is specified in the IETF Internet-Draft draft-ietf-mailmaint-aprf (revision 01 at the time of writing). A report is a JSON array of objects with a header (version, source, dkim_domain, dkim_selector, report_start, report_end, contact_info, sdi_used, extra_info) and a body of stanzas with optional segment, classification and engagement data. Reports are mailed as application/json or application/gzip attachments. Senders publish a DNS record under _aprf._domainkey to request them and can declare a Signer-Defined Identifier (SDI) header to segment the data.

Why use an APRF report analyzer?

Raw JSON with epoch timestamps and nested stanzas is hard to read and easy to misread. The analyzer converts times to UTC, lays out every segment as a table, validates the structure and makes sure nobody turns bucketed numbers into fake percentages.

Because the format is still a draft, reports from different providers may also interpret fields slightly differently. The analyzer keeps every unknown field and provider-defined category visible instead of dropping it, so nothing a provider chose to report is lost, and it tells you which draft revision the report was checked against. When you generate APRF reports yourself, the same validation shows whether your output matches the draft before you send it to anyone.

Who should use it?

Senders and ESPs receiving APRF reports, deliverability teams evaluating segment performance, and developers implementing APRF generation or processing.

How the APRF report analyzer works

Files are read in your browser: JSON directly, .json.gz decompressed with the browser's built-in gzip decoder (capped at 20 MB), and emails searched for APRF attachments. A versioned schema adapter for draft-ietf-mailmaint-aprf-01 validates the header, timestamps, SDI declaration, segments and counts. Unknown fields and provider-defined categories are kept and shown. To check the DKIM side, use the DKIM Checker.

How to use the NetTools.tech APRF Report Analyzer

  1. Open the APRF Report Analyzer.
  2. Paste the JSON, open or drop one or more .json, .json.gz or .eml files, or load an example.
  3. Click Analyze APRF Report.
  4. Read the overview, the classification and engagement tables and the segment details.
  5. With several reports, review the comparison table; export as CSV, JSON, PDF or text.

Understanding the results

Each report shows Valid, Valid with warnings, Invalid or Unsupported format, with every finding tied to a JSON path. The overview lists provider, DKIM domain and selector, the provider's version string (not the draft revision), and the UTC reporting window. Tables show counts exactly as reported; "—" means not reported. The comparison view lines reports up and warns when providers or version strings differ.

Example: a segmented newsletter report

A report for sender.example, selector mail2026, declares SDI X-Campaign-Info with separator ^ and has three segments. The weekly EU newsletter shows inbox 25,000 and unwanted 1,000. These may be bucket upper bounds, so the analyzer does not turn them into a 4% spam rate — compare segments by their reported buckets instead.

Common use cases

  • Reading daily APRF reports from a mailbox provider.
  • Comparing segments defined with an SDI header.
  • Validating reports your own system generates.
  • Checking reports from different days side by side.

Troubleshooting

Invalid JSON

The error names the line and column. Make sure the whole file was copied.

Unsupported format

The structure does not match the supported draft revision — for example a DMARC aggregate report, which is XML.

Timestamps look wrong

The draft uses epoch seconds; milliseconds are flagged.

Segments without SDI

Segments appear although sdi_used is N/A or N/F; check your DNS declaration.

Common mistakes

  • Calculating inbox percentages from bucketed counts.
  • Treating a missing category as zero.
  • Comparing reports from different providers as if definitions matched.
  • Calling APRF an RFC.
  • Reading the version field as the draft revision.

Limitations

APRF is a draft and may change; the analyzer supports draft-ietf-mailmaint-aprf-01 and names the adapter used. Engagement categories are defined by providers. Inputs are limited to 20 MB, and decompressed data to 20 MB per file.

Security & privacy

Your APRF report is processed locally in your browser. Nothing is uploaded to NetTools servers, nothing is stored, and data is cleared when you reset or leave the page.

Summary

APRF reports show where your mail lands, per DKIM domain and segment — if you read them as reported. The NetTools.tech APRF Report Analyzer validates them against the current draft, keeps every value untouched and never leaves your browser.

Frequently asked questions

Is APRF an RFC?

No. It is an IETF Internet-Draft (draft-ietf-mailmaint-aprf) in the mailmaint working group. It may change, and few mailbox providers send it yet.

Why are no percentages or scores shown?

The draft allows bucketed values, where a count is the upper bound of a range. Adding or dividing such values gives misleading numbers, so every value is shown exactly as reported.

What does “—” mean in a table?

The provider did not report that category for that segment. It is not the same as zero.

What are N/A and N/F in sdi_used?

N/A means the provider did not use the Signer-Defined Identifier you declared. N/F means no valid SDI declaration was found in DNS.

Is the report uploaded?

No. JSON parsing and gzip decompression happen in your browser. Nothing is sent to NetTools.

Can I compare reports from different providers?

They are shown side by side, but marked as not comparable: each provider defines its categories, engagement signals and bucket sizes.