What is Fraud Detection in Mobile Marketing?

Mobile fraud detection in the mobile marketing realm refers to measures taken to prevent organizations from losing money or data. Within the mobile apps ecosystem, fraud can happen in a number of ways including install hijacking, fake installs, SDK spoofing, click injection, etc.

What is Mobile Ad Fraud?

Mobile ad fraud is a type of fraud where advertisers are deceived by generating fake ad impressions, clicks, or installs on mobile devices. This results in advertisers paying for fraudulent or non-existent advertising activity, leading to a compromised data integrity and a loss of revenue. Mobile ad fraud can take many forms, such as fake apps, click farms, and bots, among others.

Why is Mobile Fraud Detection important?

Mobile fraud detection is important because it helps protect app marketers from financial losses due to fraudsters imitating real users to gain revenue. Increasing user-acquisition generally increases the probability of fraudulent activities. Preventing fraud usually requires multiple signals, which can be difficult for developers to build independently. Most MMPs have a fraud detection mechanism in place that relies on signals shared across the industry.

How is Mobile Fraud Detection done at Tenjin?

Tenjin looks at a few signals to identify fraudulent activities within apps. Some of the most common signals we use are:

• **Mean time to identify (MTTI) blocking: **This is a technique used in fraud detection to quickly identify and block fraudulent transactions by comparing the click and install timestamps. App activity is analyzed in real-time and patterns and anomalies that indicate potential fraud are identified and blocked. For example, if it’s an install hijacking fraudulent activity, fraudsters will attempt to inject a click right before an install to win the install attribution. By blocking attributions that seem unnaturally close to the install timestamp, Tenjin protects app developers from install hijacking and click injection.

• Fake Installs: This is another common type of fraudulent activity where attackers artificially inflate the number of downloads for a specific app by using automated scripts, bots, emulator devices or click farms. These can mimic the behavior of real users but usually have inconsistent device data, such as a mismatching operating system version or IP address between the click and the install. As a result, some app campaigns have a lot of fake installs that advertisers pay for and these installs do not generate any further monetary value. Identifying anomalies early prevents app marketers from incurring losses with their advertising data.

You can read more about Tenjin’s fraud dashboard and the signals we use to prevent fraud here.