APRF: A New Standard for Deliverability Feedback

There's a very exciting new proposed email specification making its way through the IETF, called APRF (Aggregate Performance Reporting). The individual draft specification comes from Alex Brotman of Comcast, Tom Corbett of Iterable, and Emil Gustafsson of Google. And Comcast is already live with a beta version of this feedback. If you want to read the full spec doc, you can find it here. But read on if you want a quick breakdown of why this is hella useful, how it works and why I'm already a big fan.

What is APRF?

Think of APRF as a simple, aggregated feedback mechanism that tells email senders how their mail is actually performing at a given mailbox provider (MBP). Much like you do with DMARC, a sender publishes a simple TXT record for their domain. In that record, you specify a reporting address (rua). If a mailbox provider supports APRF and has feedback to share, they will send you periodic (daily) aggregated reports detailing where your mail landed and how users interacted with it.

Right now, Comcast is the only mailbox provider actively sending these beta reports, but with Google co-authoring the spec, we can certainly hope for broader adoption down the line.

How to Set it Up

APRF is DKIM-based, not From domain-based. The report generator looks up the DNS record based on the DKIM signature domain (d=) and selector (s=) found in your sent mail. To set it up, you publish a TXT record at: [selector]._aprf._domainkey.[domain]. (Use an asterisk in place of the selector if you want to cover any/all DKIM selectors for that domain.)

A basic TXT record looks like this: v=APRFv1; rua=mailto:aprf-reports@example.com;

The spec also includes an optional tag called SDI (Signer-Defined Identifiers). This allows a sender to tell the mailbox provider to break down the performance data by a specific internal header you've injected (like a Campaign ID or Brand ID), up to four layers deep.

The Feedback

The data is delivered simply, in JSON format, and it's practically human readable. Here is an example of a beta report payload:

[
    {
        "header": {
            "version": 1,
            "source": "Comcast/Xfinity",
            "dkim_domain": "spamresource.com",
            "dkim_selector": "x",
            "report_start": 1783900800,
            "report_end": 1783987199,
            "contact_info": "aprf-noreply@alerts.xfinity.com",
            "extra_info": "",
            "sdi_used": " - N/A - Beta - Not in use"
        },
        "body": [
            {
                "classification": {
                    "inbox": 100,
                    "unwanted": 0
                },
                "engagement": {
                    "positive": 0,
                    "negative": 0
                }
            }
        ]
    }
]

Why This is Cool

We here in the land of deliverability are not strangers to inbox placement tracking. Platforms like Inbox Monster and Validity's Everest have long provided placement data. However, those platforms traditionally rely on seedlist testing (sending mail to a small collection of test mailboxes) to estimate how your mail is treated. Don't get me wrong; those platforms still have much to offer. But, APRF is something different. It shares real, aggregate data driven by actual user behavior.

The report gives you two vital buckets of data:

• Classification: How many of your messages actually went to the inbox versus how many went to the unwanted(spam) folder.
• Engagement: It tracks user votes. You get counts for positive actions (like a user pulling your mail out of the spam folder or marking it as safe) and negative actions (like standard spam complaints unsubscribing, or deleting without reading).

My take is that this is incredibly valuable insight that will easily help a good sender be a better sender.

TL;DR? I Like

This is just so simple. So easy. If I were building a deliverability dashboard or an email metrics portal today, I'd be adding APRF support immediately. Even without a fancy dashboard, just having access to this lightweight, easy-to-digest data directly from the mailbox providers is hella useful. In short: I'm loving this. I hope to see it gain wide industry adoption and much success. Excellent work, Alex, Tom, and Emil!