Maintain two counts: documented eligible content and observed usable content. Do not extrapolate a successful sample into a claim that the whole catalogue works.
Use this waterfall:
Advertised records → distinct titles/configurations → buyer-relevant content → documented eligible content → tested content → observed usable content.
At every step, retain a separate unknown bucket. Apply exclusions in a stated sequence so overlapping issues are not subtracted twice.
| Catalogue measure | Recommended definition |
|---|---|
| Documented eligibility | Relevant title/configuration with sufficient permission evidence for the specified audience and property |
| Observed usability | Eligible title/configuration that passes the defined launch, interaction and presentation checks in required tested environments |
| Weighted requirement coverage | Sum of priorities for satisfied buyer requirements ÷ sum of priorities for all buyer requirements |
| Essential-title coverage | Passed must-have requirements ÷ all must-have requirements; proposed approval target: 100% |
| Untested eligible coverage | Eligible content without sufficient operational testing; report separately, not as usable |
First Look Games’ August 2026 guidance says its demo session counter measures page loads, not distinct visitors or completed plays, and its dashboard does not provide per-game play counts. This is a specific example of why a supplier dashboard’s metric name needs a written definition before it becomes procurement evidence. (firstlookgames.com)