Google App Ads: 60% of Installs Were Bots
Original: I spent $220 on Google app ads and 60% of the installs were robots
Why This Matters
Bot farms actively exploit Google's conversion optimization loops, quietly draining small developer budgets at scale.
Nick Abe, developer of puzzle app Dayzle, spent CA$220 on Google Ads targeting Android installs. Google reported 56 installs over two weeks; 33 showed bot-farm patterns, 7 came from untargeted countries, and only 13 appeared to be real users — who completed 92 games combined.
Abe set up a Google Ads campaign at CA$40/day with an install goal and a CA$1.50 target CPA. When Google couldn't spend at that price, he removed the target. Google immediately spent CA$80 in one day and reported 21 installs. His own admin panel showed 1. Digging into raw analytics, he found 20 new Android devices running an app version the Play Store had stopped serving days earlier — meaning those devices got the APK from somewhere else, yet all reported Google Play as the installer. Each device opened the app once, spent zero seconds on any screen, and never returned. Twenty-eight distinct phone models across nineteen U.S. states — a suspicious spread for 20 devices that behaved identically. Over the full two-week campaign: 56 billed installs, 33 matching the bot pattern, 7 from countries outside the campaign's targeting, and 13 apparent real users. The mechanics are straightforward: the farm watches the shortest ad video (no click needed), then installs a cached APK — faster than the Play Store and less likely to trigger detection. Google counts a view followed by an install as a conversion, so the farm's activity actually improved the campaign's apparent performance, directing more ad spend toward itself. Abe has filed an invalid-traffic refund request with Google and changed the campaign goal to 'won a puzzle' — reasoning that scripting a Sudoku solve is harder than faking an open.