COLLABORATIVE CLICK FRAUD DETECTION AND PREVENTION SYSTEM (CCFDP) IMPROVES MONITORING OF SOFTWARE-BASED CLICK FRAUD Li Ge, Darren King Hosting.com 501 S. 4 th ST. Louisville, KY 40202, USA Mehmed Kantardzic CECS Department University of Louisville Louisville, KY 40292 USA ABSTRACT Click fraud had been identified as biggest threat to Pay-Per-Click advertising business model. We analyzed different types of click fraud activities and proposed a new classification of click frauds into four categories. While traditional commercial approached detect only some specific types of a click fraud, we developed a new system to detect and prevent all four major click fraud categories. CCFDP system is based on the collaboration between server side log and client side log. Our approach assumes that: a) user detail activities inside a web page differentiate a normal user from human click fraud while b) server side log can reveal robotic software click fraud by analyzing the difference between those two logs. Proposed architecture of the CCFDP system and click fraud identification algorithms are present in the paper. Preliminary experimental results are also included. KEYWORDS Click Fraud, Detection, Prevention, Software Click 1. INTRODUCTION Pay-per-click (PPC) is an online advertising payment model, used by search engine companies, in which payment is based solely on qualified click-troughs. This pay-per-click model is now the fastest-growing form of internet advertising, according to the Interactive Advertising Bureau. However, the cost for pay-per-click becomes very high, varying by keywords and list position. Some businesses pay Google or Yahoo's Overture $90 per a click to appear as the No. 1 or 2 ads, while at the same time companies report that their fraudulent traffic is higher than 50% and the losses are in the range of $5,000 to $300,000 (Lycos, Inc.). Click Fraud is a scam involving setting up a website affiliated with a major search engine, displaying pay- per-click advertising from the search engine and then using various methods to fraudulently increase the number of clicks to the advertiser from the affiliate website (Metwally, A. et al., 2005). The affiliate website receives a portion of the money generated by the click through even though the clicks were not generated by genuine customers. It was identified to be one of the biggest threats to the internet economy. Fig 1 shows two click fraud examples. Figure 1(a) is the human click fraud. People click on an advertiser’s PPC links from client computer to navigate to advertiser’s web site multiple times without view the contents of the site. Fig. 1(b) shows some software can click the advertiser’s link instead. The approach to click fraud analysis and detection has some similarity to web log user behavior pattern research. Many researches had been carried out for user’s activity based on web logs (Banerjee 2001, Berendt 2000, Shahabi 1997, Gehrke 2001). However, the user activity pattern and page link analysis based on web ISBN: 972-8924-06-2 © 2005 IADIS 34