What Is a Feedback Loop (FBL)?
TL;DR
A feedback loop, or FBL, is a service where an inbox provider tells a sender that a recipient marked one of their messages as spam. Most providers report in the Abuse Reporting Format defined by RFC 5965, so the complaining address can be suppressed automatically, and enrolment is keyed on either the DKIM signing domain or the sending IP.
Which providers offer a feedback loop, and on what key?
Yahoo runs a Complaint Feedback Loop that supports DKIM-signed mail only, so enrolment is per signing domain and the reports arrive in Abuse Reporting Format. Microsoft splits it in two: the Junk Mail Reporting Program sends per-complaint reports and is keyed on sending IP addresses you register, while Smart Network Data Services gives an IP-level dashboard.
Google works differently and the difference trips people up. The Postmaster Tools feedback loop is keyed on a Feedback-ID header you add and DKIM-sign, and it reports aggregate spam rates per identifier for gmail.com recipients rather than naming individual complainers, with data appearing only when there is enough traffic and enough distinct reports to aggregate safely.
What do you do with feedback loop data?
Suppress on arrival, permanently and without a second thought. Someone who pressed the spam button has told two parties something: you, and the provider whose filter decides your placement, and the second one is why re-sending to that address is the most expensive mistake available.
Then read the attribution. Because the Gmail loop groups by an identifier you choose, setting the Feedback-ID to something meaningful, such as a campaign, turns an aggregate rate into a diagnosis of which message caused it.
Frequently asked questions
Partly. The IP-keyed programmes report individual complaints at any volume, but the Gmail loop only surfaces data once there is enough traffic and enough distinct reports to aggregate, so a small sender can be causing complaints that never appear in a dashboard.