Definition
The share of algorithmically generated leads whose eventual case outcome is returned to the generating system as labeled ground truth.
The Measurement Gap
There is no metric that tells you whether case outcomes return to the systems that generated the leads, which means deployed accuracy is frozen at its launch state while being described as learning.
Formula
LDFR = L_fedback / L_outcome, with median lag from disposition to write-back.
Proposed formula, pending validation pilots.
At a Glance
Decision Supported
Whether the system can improve from its own operational record, or is permanently frozen at deployment accuracy.