How to Detect Fake App Installs

Install fraud  publishers, ad networks and marketers, resulting in lost ad spend and skewed data. It can be caused by bots, incentivized downloads and SDK spoofing. The fraud players can be divided into two categories: ad networks that sell fake installs and fraudulent publishers who engage in click injection or attribution fraud.

In most cases, install fraud is executed by bots. These devices run code on the background of a user’s phone without their knowledge to mimic human behavior, including clicking on ads and installing apps. They can also trigger in-app events and even fake post-install activity. They can be controlled by a central master or by users themselves who give them commands. This type of fraud can be detected by comparing traffic from different sources or by looking for unusual device activity like the device’s movement (via its accelerometer and gyroscope), battery data, or user-screen interaction.

Mastering the Art of Safeguarding Your Ad Campaigns: Detecting and Mitigating Invalid Traffic

Another way to detect fake app installs is by focusing on the stage of the mobile marketing funnel where you pay for your installs, i.e. when you buy CPI traffic, check for impression fraud; if you pay per app install, check for install fraud. A high percentage of new devices coming from a single publisher can indicate that there is suspicious activity. You can also compare your MMP ad tracker’s data with download numbers from the app stores. If the numbers differ significantly you can suspect that there is fraud in your campaign.