Mediation update diagnosis
Fill rate dropped after a mediation update
A fill-rate drop after a mediation change is a timing signal, not proof that the update caused it. Compare the same segments before and after the rollout, then inspect what changed in serving behavior.
Confirm the rollout boundary
Record the app version, SDK version, mediation configuration, and rollout time. A change that began before the rollout needs a different explanation.
Keep match rate and fill rate separate
Match rate and fill rate use platform-specific definitions. Compare requests, matched requests when available, fills, and impressions without assuming that one rate proves the behavior of the other.
Check the source and segment level
A global average can hide a source, country, format, or ad-unit failure. Look for the smallest segment that changed immediately after the update.
Post-update checks
- Compare the same country, placement, and ad source before and after the rollout.
- Check requests, matched requests when available, fills, impressions, and show rate separately.
- Verify SDK and adapter versions, initialization, consent, and error logging.
- Review source availability, bidding or waterfall configuration, and timeout behavior.
- Check price-floor and instance changes made near the update.
- Avoid changing multiple configuration items until the affected segment is confirmed.
How to read it
Common interpretation
If requests are stable but fills fall in the newly released segments, the update, a source configuration, or a serving condition deserves investigation. If the movement is broad across unchanged versions too, demand or traffic may be the stronger explanation.