Impressions stable, weighted eCPM down
- Observed change
- Exposure stays close to normal while rewarded pricing falls.
- Most likely driver
- Pricing or demand movement after mix checks
Mobile ad revenue drop diagnosis
Find what most likely caused your mobile ad revenue drop.
Use anonymized CSV rows to separate impressions, fill rate, country mix, placement, ad source, time of day, and weighted eCPM before changing floors or mediation settings.
See sample CSVBrowser-only CSV parsingNo SDKNo dashboard loginAnonymized rows
01Diagnose
Seven signals before you blame eCPM.
Start with the signal that moved first. A blended number is a headline, not a diagnosis.
Traffic
Traffic and exposure
Impressions and impressions per active user show whether exposure dropped before pricing did.
- Typical anomaly
- Revenue fell while eCPM stayed normal because fewer sessions reached the rewarded placement.
- Next check
- Compare DAU, sessions, requests, and impressions per session.
02Separate
A short evidence trail, not another dashboard.
The public workflow keeps the first decision simple: compare matching periods, find the strongest signal, then decide what deserves a deeper check.
- 01
Upload
Use sample data or an anonymized CSV.
- 02
Separate
Compare traffic, fill, GEO, placement, source, and pricing.
- 03
Act
Get a likely driver, supporting signals, and first checks.
03Patterns
Four common ways a revenue drop is misread.
Every chart below is a directional sample, not customer data. The case pages explain the evidence to look for.
Fill rate dropped before revenue
- Observed change
- Requests remain present, but fewer opportunities are filled.
- Most likely driver
- Source availability or mediation serving
Country mix distorts blended eCPM
- Observed change
- Top-country pricing is stable; lower-value GEO share grows.
- Most likely driver
- Traffic composition, not a global price collapse
Live events reduce peak-hour impressions
- Observed change
- Peak exposure falls while price and fill stay close to normal.
- Most likely driver
- Audience timing or product behavior
04Fit
Built for the first investigation.
Start with the diagnosis. Escalate only when you have a clearer evidence trail for a platform, mediation partner, or deeper review.
Best for
- Live mobile apps or games with ad revenue
- Two comparable reporting periods
- AdMob, MAX, LevelPlay, TopOn, or custom exports
- Teams able to share anonymized CSV rows
- Questions about traffic, fill, GEO, placement, source, timing, or pricing
Not designed for
- Account approval or policy appeals
- Apps without meaningful ad traffic
- User-level or personal-data analysis
- Guaranteed revenue recovery
- Automatic mediation or price-floor changes
05Safety and method
Your account stays out of the first diagnosis.
The public demo parses CSV locally in the browser. Remove private identifiers before sharing anything for a second look.
- No account credentials
- No API keys
- No SDK
- Remove private identifiers
- CSV stays in the browser
1Anonymized CSV
2Browser parser
3Diagnosis result
06Resources
Use the right level of help for the question.
Begin with a sample diagnosis, then move into templates, cases, and method notes only when you need more context.
eCPM Bazaar
Not sure what changed?
Try the sample diagnosis first. Send anonymized rows only when you need a second look.