Bots faked 77% of a developer's Google app installs

Google billed a developer CA$220 for two weeks of App Campaigns ads. The app's own admin panel logged only 13 real users against Google's count of 56 installs. That 43-install gap is what sent the developer digging through device logs instead of trusting the dashboard.

Twenty of the devices were running an app version no longer available through the Play Store. That build turned up on 28 phone models spread across 19 states. Each device opened the app once, spent zero seconds on screen, and never opened it again.

The developer tallied 33 of the installs as bot traffic and 7 more as mis-targeted rather than fraudulent. Together that puts about 77% of the 56 billed installs at zero value, per Dayzle's breakdown.

This is one small campaign, not a platform-wide measurement, and the post doesn't claim otherwise. But the diagnostic holds at any scale. Installs billed by an ad platform and active users logged by an app's own backend should track each other. A device running an app version that's no longer distributed is not something a real user does.

That invalid-traffic form is filed. No refund, no confirmation, no timeline has followed.