The anatomy of mobile ad fraud: Intelligent Device Farms
While Device Farms are losing its share on app install fraud market, a more advanced type of such farms is gaining momentum. This type of fraud is significantly different from the classic Device Farms: fraudsters here perfectly fake post-install activity, engagement, and even payments. At Scalarr we call this Intelligent Device Farms and consider them a part of the New Face of Mobile Fraud category.
1. In 2018 Intelligent Device Farms were responsible for 6,1% of all fraud cases.
2. Social Casino and eCommerce categories are most vulnerable to this type of fraud due to their high CPI rate (from $3 to $10).
3. Intelligent Device Farms have a wide range of devices and other attributes and can emulate the real user's behavior for 14 days or even more, remaining undetected.
How Intelligent Device Farms work
1. Intelligent Device Farms use a wide range of devices and install attributes to avoid obvious signs of fraud (device ID duplicates, IP-addresses duplicates, etc.).
2. Fraudsters trying to connect simultaneously to many publishers (including conditional trusts) and split their volume into many small clusters.
3. Finally, Intelligent Device Farms perform all the needed KPIs / events for good traffic quality.
The close similarity of Intelligent Device Farms to real users is achieved in several ways:
1) These can be device farms managed by real people (depending on the size of the farm, the number of people varies), that perform all the required events manually.
2) Some of the Intelligent Device Farms use a special program that runs processes on all devices, thus simulating the behavior of real users.
How to deal with Intelligent Device Farms
The classic Device Farms can be identified by numerous installs from one device and by numerous installs from several identical IP addresses. But since these marks were the most obvious sign of ad fraud, Intelligent Device Farms use more advanced approach, constantly changing IPs, device IDs and applying automated tools for various actions. This type of fraud is very dynamic with constant reverse engineering, and its successful detection requires a detailed analysis of post-install events using Machine Learning and Big Data technologies.
Read Scalarr’s Ultimate Guide to Mobile Fraud Types to understand more about dealing with Intelligent Device Farms in all specific cases.
The general principle of “mixes” grounds on the conscious use of several various types of fraud to get over the known protection measures of fraud ...
Fraud is an adaptive crime, so it needs special methods of intelligent data analysis to detect and prevent it