Four common ways mobile ad revenue drops can be misread.
Before sending your own data, review how eCPM Bazaar turns anonymized mobile ad monetization rows into a diagnosis card, most likely drivers, and recommended next tests.
Scenario
Case 1: eCPM dropped, traffic stayed stable
Observed change
Revenue fell 24%. Impressions were almost flat, but weighted eCPM dropped from $18.40 to $13.90.
Why this interpretation fits
This looks more like a pricing or demand issue than a traffic issue. The first checks should be country mix, bidder/source performance, eCPM floors, seasonality, and recent mediation changes.
Supporting signals
Traffic volume stayed stable, so the most likely driver is pricing, demand, source performance, floor settings, or country-level eCPM changes.
Directional exampleeCPM Bazaar
Revenue dropped 24%
Diagnosis statusDirectional example
Most likely drivereCPM dropped
Severity
Medium
Country
Mixed geos
Placement
Rewarded + interstitial
Ad source
Top demand sources
Compare eCPM by country and source before changing traffic or placement logic.
Takeaway
Do not change traffic or placement logic before checking whether pricing changed.
What to avoid
Avoid treating a pricing drop as a traffic-quality problem before checking country-level eCPM and source performance.
Diagnosis logic
When impressions stay stable but revenue falls, the first suspect is weighted eCPM, source demand, floor settings, or country-level pricing.
Data limitations
This is a directional example. A real diagnosis should compare the same segments across matching periods.
Recommended checks
Compare eCPM by countryCheck top ad sourcesReview floor or waterfall changesLook for platform status or demand changes
Revenue fell 31%. eCPM stayed close to normal, but fill rate moved from 78% to 54%.
Why this interpretation fits
The eCPM number alone is misleading here. The most likely driver is fill or match pressure, so changing price floors blindly could make the diagnosis worse.
Supporting signals
Demand price may still be normal, but fewer requests are turning into filled impressions.
Directional exampleeCPM Bazaar
Revenue dropped 31%
Diagnosis statusDirectional example
Most likely driverFill rate dropped
Severity
High
Country
United States
Placement
Rewarded Video
Ad source
Unity Ads
Check source availability, timeout, mediation release, and floor settings first.
Takeaway
If fill rate drops while eCPM stays stable, blindly adjusting price floors can make the diagnosis worse.
What to avoid
Avoid changing eCPM floors first when the real issue may be fewer requests turning into filled impressions.
Diagnosis logic
Stable eCPM with lower fill rate usually points to source availability, timeout, SDK, mediation, or request logic issues.
Data limitations
This is a directional example. A real diagnosis should compare the same segments across matching periods.
Recommended checks
Split requests and fills by countryCheck timeout and request logicReview source availabilityInspect recent SDK or mediation releases
Total eCPM dropped 18%, but US eCPM was stable. More impressions came from lower-eCPM countries.
Why this interpretation fits
This is probably a traffic mix issue, not a global pricing collapse. The team should compare revenue contribution by country before changing global mediation settings.
Supporting signals
Blended eCPM can fall even when top-country eCPM stays stable, if more impressions come from lower-eCPM countries.
Directional exampleeCPM Bazaar
Total eCPM dropped 18%
Diagnosis statusDirectional example
Most likely driverCountry mix changed
Severity
Low
Country
US stable, BR/IN share up
Placement
All placements
Ad source
All sources
Separate country mix from source pricing before changing global mediation settings.
Takeaway
Do not treat a traffic mix shift as a global demand collapse.
What to avoid
Avoid changing global mediation settings before separating Tier 1 trends from blended traffic mix changes.
Diagnosis logic
Blended eCPM can fall even when top-country eCPM stays stable if more impressions shift toward lower-eCPM regions.
Data limitations
This is a directional example. A real diagnosis should compare the same segments across matching periods.
Recommended checks
Compare impressions share by countryCheck UA campaigns and organic traffic sourcesReview placement exposure by regionSeparate country mix from source performance
Revenue looked weaker during the week, but eCPM and match rate stayed mostly stable. The real drop came from fewer impressions during the app's normal evening peak window.
Why this interpretation fits
This looks less like an ad demand problem and more like an audience behavior shift. Daily averages can make the revenue trend look worse than the ad stack really is.
Supporting signals
Sports events, holidays, exams, work schedules, or local events can reduce user activity during the app's normal peak monetization window.
Directional exampleeCPM Bazaar
Revenue dropped 19%
Diagnosis statusDirectional example
Most likely driverPeak-hour impressions dropped
Severity
Medium
Country
Sports-heavy GEO
Placement
Native + interstitial
Ad source
All sources
Compare hourly impressions and impressions per user before changing mediation settings.
Takeaway
If revenue drops during the app's normal peak hours, split by hour before assuming eCPM, demand, or mediation failed.
What to avoid
Avoid assuming the ad stack failed when impressions per user drops only during event hours or normal peak windows.
Diagnosis logic
If eCPM and match rate stay stable while peak-hour impressions fall, the most likely driver is audience behavior rather than ad demand.
Data limitations
This is a directional example. A real diagnosis should compare the same segments across matching periods.
Recommended checks
Compare event days vs normal daysCheck the normal peak window separatelyCompare event hours vs nearby non-event hoursCheck impressions per userConfirm eCPM and match rate stayed stable