Time-of-day diagnosis
Why AdMob revenue can drop during live events even when eCPM is normal
A revenue drop can look like a monetization problem at first. But if impressions fall during the app's normal peak hours while eCPM and match rate stay stable, the issue may be user behavior, not ad demand.
Why daily revenue can be misleading
Daily totals hide when the movement happened. A short drop during the app's strongest monetization window can reduce revenue even if the rest of the day looks normal.
Check hourly impressions before changing mediation
Before changing floors, SDKs, or ad source settings, compare hourly impressions and impressions per user. If the drop is concentrated in a peak window, the ad stack may not be the first suspect.
Compare event days vs normal days
Sports matches, holidays, exams, work schedules, local events, and seasonal behavior can shift audience attention. Compare event days with nearby normal days, not only this week against last week.
Separate impressions per user from total impressions
Total impressions can fall because fewer users visited, or because active users saw fewer ad moments. Impressions per user helps separate traffic volume from session behavior.
When it is probably not an eCPM problem
If eCPM, match rate, and fill rate stay mostly stable while peak-hour impressions drop, the most likely driver is audience behavior rather than demand collapse.
Try a browser-only CSV diagnosis
Use anonymized rows to compare affected hours, normal hours, country mix, placement, ad source, and weighted eCPM before deciding what to change.
Practical checklist
- Compare normal peak hours vs affected hours.
- Compare event days vs non-event days.
- Check impressions per user.
- Check match rate / fill rate.
- Check eCPM by country and placement.
- Only then decide whether demand or mediation changed.
How to read it
Common interpretation
If revenue dropped during peak hours but eCPM and match rate stayed stable, the first explanation to test is audience timing. Daily averages can make a behavior shift look like an ad demand problem.