Monday, 20 November 2017

Here's How Mobile App Publishers Are Defending Against Ad Fraud

Blueprint includes a mix of rule-based and statistical methods

The digital ad industry has had something of a rough year, with concerns emerging—and re-emerging—about brand safety, ad transparency, viewability and fraud.
A new report by Singular examines some of the top methods used by fraudsters to manipulate the mobile ad ecosystem, as well as some effective defenses against such efforts.
The report broke down fraud into two main categories. One is fake users, in which dishonest players rely on a mix of bots, malware and install farms to generate activity by means other than real people. Some techniques in this category include install farms that rely on humans to install and interact with apps to generate the appearance of a large audience and mobile device emulators that mimic the effect a large number of real mobile devices might have.
According to Singular’s breakdown, the second category of mobile ad fraud is attribution manipulation, in which fraudulent clicks are sent to an attribution record system, thereby falsely taking credit for an engagement—and gaming last-click attribution models in the process.
This can include click injection—when fraudulent apps are downloaded by users and generate fake clicks to take credit for the installation of other apps. It also includes click spamming, when a fraudster uses real, but appropriated, mobile IDs to send out a host of fake click reports. When a real user with that ID organically installs an app, the fake click will get the credit.
Methods Used to Prevent Mobile Ad Fraud Among Mobile Apps Worldwide, Sep 14-Oct 14, 2017 (% of total)
But those running mobile campaigns can rely on a few different types of fraud prevention using both rule-based and statistical methods. Singular’s analysis of its mobile app network worldwide from September 14 to October 14, 2017, found that time-to-install (TTI) anomalies were used to prevent 36.1% of mobile ad fraud, followed by geographic outliers (20.9%) and IP blacklisting (18.6%).
TTI anomalies are fraudulent traffic identified using a statistical analysis to reveal abnormal behavior that correlates to click spamming, while geographic outliers identify a large distance between the location of a click and the corresponding app download. Meanwhile, IP blacklisting identifies IP addresses that don’t belong to actual users, or that had other signs of fraud.
While there are solutions to mobile ad fraud, Singular also cautioned that fraudsters and those combating fraud are stuck in a never-ending game of cat and mouse. As soon as new methods to combat ad fraud are developed, it’s almost a certainty that less ethically inclined people will attempt to find workarounds.