Fraud Scanner

Fraud Scanner protects your data against different types of fraud, such as click injection, impressions flooding, fake devices, and others that distort the results of ad campaigns and destroy the budget.

With advertising costs climbing, this industry attracts more and more scammers. Rogue players may come in all shapes and sizes, be it your users, trusted advertising agencies, or even seemingly reliable partners.

To protect your data and not pay for any fraudulent traffic, you can track fraud metrics with MyTracker.

How it works

Fraud Scanner is available for mobile and web traffic.

Every day, the system monitors incoming traffic and calculates fraud metrics for each install and site visit, e.g. how many installs come from suspicious devices, what is the share quick installs, and so on.

Device fraud metrics are updated every 24 hours during the first 3 days as new data flown in. The final and most accurate result of the fraud monitoring is fixed 3 days after an install or a site visit.

It only tracks the fraud metrics, without blocking the traffic automatically. This way, you can make an informed decision (for example, to fine or disconnect the partner)

Stages

Fraud monitoring is done in several stages:

1. Hardware stage. At this stage, the system recognizes fraud based on hardware properties. For example, if an Android app is installed on a device running iOS, or an app is being launched from different countries over a short period.
2. Click stage. Here, the antifraud mechanism comes into play when a user interacts with your ads and remains active until the activity on the app or website. The system will identify all dubious clicks and impressions: quick installs, low CCR, and etc.
3. In-app stage. This final stage kicks in after an app is installed or a site visit. The system will detect abnormal behavior, such as not launching the app or having only short sessions with it.

Based on the results of all the stages, Fraud Scanner will put different fraud metrics together to give you a sense of the general traffic quality (strict, confident, or soft fraud). These metrics include deferred installs and a low CR and can be seen individually.

Benchmarks

To properly assess the threat of fraud, you can use special indicators called benchmarks. These are reference values which, if exceeded, cause the corresponding indicators to get highlighted in red, warning you about potential fraud.

The benchmarks have been calculated separately for each fraud metric based on machine learning and a large pool of data for iOS, Android, and Web.

For example, the Android benchmark for Suspicious devices stands at 0.69%. If more than 0.69% of installs come from suspicious devices, MyTracker will highlight this value in red so you can promptly detect the fraud. For more details, refer to the How to read reports section.

How to use

We recommend regular fraud monitoring:

1. Analyze the Fraud Scanner report.
2. Build the detailed report using Builder.
3. If fraud is detected, you can notify a partner and take action to prevent fraudulent attacks (for example, fine or disconnect the partner).

For more details, refer to sections Fraud detection and Fraud reconciliation with partner

The available fraud metrics, templates, and traffic details vary between the Basic and Premium plans.

Basic plan

The basic functionality of Fraud Scanner allows you to detect fraudulent traffic using Combined metrics and use Artificial Intelligence Fraud Detector to monitor fraud.

The basic plan is enough to assess the threat of fraud and see if there is a need to subscribe to the Premium plan.

Premium plan

The Premium plan unlocks all Fraud Scanner capabilities:

• sort the stats by Partner, Campaign, SubID, Traffic type, Country, and identify the sources of fraudulent traffic;
• analyze a variety of metrics (Suspicious Devices, Quick Installs, Post-install fraud, etc.) to identify fraudulent schemes;
• refer to detailed templates (Strict, Confident, or Soft fraud) to evaluate your traffic easily and accurately.

To switch to the Premium plan, please contact your MyTracker manager or send an email to support@tracker.my.com

Final recommendations

• Do not work with partners that have any history of fraud
• Make a point of knowing your partner's traffic in the most detail possible
• Take note of the slightest changes in fraud metrics
• Configure event tracking so as to easily trace conversion funnels. Apart from target events, which are easy to fake, track simple in-app events, such as, say, beating level 5 or 10
• Stay alert even with a trusted partner.