When our website or web application suddenly receives visits from locations outside the average customer base, or registered users provide data that has a pattern, it may mean that we have been targeted by fraudulent traffic. Detecting this kind of activities is important for many reasons. In the online advertising industry, fraudulent traffic causes additional costs for every market participants. Publishers have to pay for infrastructure used to service. Advertisers pay for traffic that has no value. Other examples are payments processors, all sorts of contests, betting and online casinos.
Because of this, there is a need for sophisticated fraud detection systems. Detecting and preventing traffic fraud to maximize profits and minimizing operational costs is very challenging for publishers.
In this article, I show you how to detect suspicious visits to your website. We will check if IP geolocation and other IP-related characteristics could differentiate good and bad traffic. In the end, we will build rules that help automate the process of fraud detection.