Definition
The share of served predictions or generations that receive a returned outcome signal usable for evaluation or training, within a defined return window.
The Measurement Gap
There is no metric that tells you what share of served predictions ever receive a returned outcome signal, which means the platform cannot learn which models are good.
Formula
OSRR = served_items_with_returned_outcome / served_items, by model and consumer
Proposed formula, pending validation pilots.
At a Glance
Decision Supported
Which models the platform can actually learn about, and where outcome instrumentation investment goes next.