A tangent about Google‘s Feedback Loop

Google‘s anonymized, aggregated, nontraditional feedback loop feature allows senders to include a special Feedback-ID header in their mailings.

When implemented, this header allows you to attach up to three¹ higher-level tags to each mailing you send. These tags can be anything you want to classify, like a type of mail or an audience segment.

In a dedicated ”Feedback Loop“ screen, Google Postmaster Tools will then give you insights into whether some tags (which they call Identifiers) are seeing higher rates of spam complaints than others.

As always with Google Postmaster Tools, the exact mechanisms here are not well documented, so you may feel somewhat out of control.

For example, a tag will not appear on your Postmaster Tools reports if it has a low absolute number of occurrences in your daily spam complaints. In part this is to prevent senders from (mis)using tags as a way of tracing complaints back to individual recipients.

This also means that if you are not a very high volume sender, fine-grained identifiers like specific mailing/campaign IDs may also not be very useful — if you‘re sending 1,000 – 100,000 mails per day, you almost certainly won‘t get enough spam complaints per individual campaign to see anything here.

There are some oddities in the documentation and in how this is implemented by sending software (when it is implemented at all) — in particular, our sense from the documentation is that this is geared toward large-volume vendors who send mail on behalf of many different customers using a shared domain. Those vendors would be able to use the Feedback Loop Identifiers to track down which of their customers are triggering spam complaints, potentially impacting the whole sending domain.

But in theory you could use this for a wide variety of interesting classification approaches, using the three tag dimensions to:

• classify and analyze spam complaint patterns across your primary bulk mailer versus secondary tools like Salesforce and Quorum
• track spam performance on transactional after-action mailings versus bulk mailings
• compare your day-to-day mail stream against messages sent via automations like your welcome series, reengagement series, or event reminders
• classify donation asks versus advocacy asks
• or look at different list segments.

At the moment we’re not aware of any advocacy toolkits, bulk mailing platforms, or email marketing software that give you the ability to control this Feedback-ID header and take full advantage of the Postmaster Tools feature, but we think it would be very neat!

ActionKit

ActionKit implements this header but doesn‘t allow you to change it. Their header includes the mailing ID (rightmost component), some sort of stable client instance identifier (leftmost component), and … the letter ”b”, always, in the middle. (The “b” is for “bulk,” as in “not transactional.”)

Action Network

Action Network doesn‘t currently support this header.

Engaging Networks

Engaging Networks implements this header but doesn‘t allow you to change it. Their header includes two identifier components (instead of three) — the rightmost is some sort of numeric identifier that seems to stably identify a client instance, and the leftmost is presumably an email campaign identifier.

EveryAction

EveryAction doesn‘t currently support this header.

Hubspot

Hubspot doesn‘t implement this header for you, and doesn‘t provide any way for you to implement it yourself, so this is unsupported.

Luminate

Luminate doesn‘t currently support this header.

Mailchimp

Mailchimp implements this header but doesn‘t allow you to change it. Their header includes the Mailchimp data center name in the rightmost component, and two other numeric information that presumably identify a Mailchimp customer and email campaign.

Salesforce Marketing Cloud

Salesforce Marketing Cloud implements this header but doesn‘t allow you to change it. Their implementation is documented and includes the message‘s Marketing Cloud account ID, JobID, and sending IP address.