Google is making a "reasonable" effort to fight click fraud, according to a report written by a New York University computer science expert.
Google commissioned the 47-page report as part of a class action lawsuit settlement reached last March, according to Shuman Ghosemajumder, a business product manager with Google. The report was written by Alexander Tuzhilin, a professor of information systems at New York University.
"We agreed with the plaintiffs to have an independent expert examine our detection methods, policies, practices, and procedures and make a determination of whether or not we had implemented reasonable measures to protect all of our advertisers," Ghosemajumder wrote Friday on Google's official blog. His post can be found here: http://googleblog.blogspot.com/2006/07/findings-on-invalid-clicks.html
Google and several other pay-per-click providers, including Yahoo and America Online, were sued by advertisers who claimed they were charged for invalid clicks dating back to 2002. Google has agreed to pay as much as $US90 million to settle the case.
The practice of click fraud -- clicking on ads to cost a competitor money or to bump up commissions for Web publishers -- has come under scrutiny of late, and some experts worry that it ultimately could threaten Google's main source of revenue, pay-per-click advertising.
Google reported revenue of $US6.14 billion for fiscal 2005, most of which came from pay-per-click ads.
After meeting with Google engineers over the course of three visits and examining internal documents, Tuzhilin concluded that Google's Click Quality team, formed in 2003, is now doing a reasonable job in fighting fraud.
"Although the Click Quality team's solutions were still not perfect ... the invalid clicking problem at Google was 'under control' by the end of 2005," Tuzhilin wrote.
In fact, Google was hit by several "massive" click fraud attacks in 2005, Tuzhilin wrote. "Although not perfect, Google detection software managed to remove massive amounts of clicks during these attacks.'
Tuzhilin's report can be found here: http://googleblog.blogspot.com/pdf/Tuzhilin_Report.pdf