SES NY: Click Fraud Remains A Contentious Issue
Yesterday, Click Forenics released its quarterly Click Fraud Index showing a purported increase in click fraud from 14.2 percent in Q4 ’06 to 14.8 percent in Q1. By contrast, Google argues that “click fraud” constitutes less than 0.02 percent of all clicks. Chris Sherman earlier summarized the Click Forensics findings. Click Forensics’ CEO Tom Cuthbert […]
Yesterday, Click Forenics released its quarterly Click Fraud Index showing a purported increase in click fraud from 14.2 percent in Q4 ’06 to 14.8 percent in Q1. By contrast, Google argues that “click fraud” constitutes less than 0.02 percent of all clicks. Chris Sherman earlier summarized the Click Forensics findings.
Click Forensics’ CEO Tom Cuthbert was on the “Auditing Paid Listings & Click Fraud Issues” panel at SES last week, which was one of the most interesting and contentious panels I attended at the conference. Also on the panel were Click Tracks CEO John Marshall, Google’s Shuman Ghosemajumder and Yahoo’s new VP of Marketplace Quality Reggie Davis.
Marshall took the position that it was often difficult to distinguish click fraud from poorly performing ads and used a case study to illustrate the point. He said that catching click fraud will necessarily involve a combination of machine detection and human scrutiny and that no automated system by itself will be fool proof. Marshall explained that hallmarks of click fraud include ads with low conversion rates, lots of single-page visits and lots of clicks from countries where marketers are unlikely to have potential customers or prospects. He added that no fraud event will likely trigger all of these but every instance of click fraud will likely have at least one of these indicators.
Google’s Ghosemajumder outlined Google’s worldview regarding click fraud and invalid clicks (he generally avoided the term “click fraud” in favor of the more neutral “invalid clicks”). He said the two motives for click fraud are “attacking advertisers” and “inflating revenue.” He then discussed Google’s various internal efforts to detect invalid clicks. He explained that Google doesn’t charge advertisers for clicks it deems invalid and that there are millions of dollars lost to Google through this process. Accordingly, he disputed the argument some have advanced that search engines benefit from click fraud and are thus weak on enforcement or in their efforts to prevent it. He wouldn’t go into great detail about some of Google’s methods because he said he “didn’t want to educate fraudsters.”
Click Forensics’ Tom Cuthbert was the third speaker and laid out his company’s Click Fraud Index data, which was dramatically at odds with the Google statements regarding the incidence of Click Fraud. Cuthbert vigorously defended the role of “third parties” (i.e., Click Forensics) vs. search engines and/or the IAB in policing and reporting on the issue.
Yahoo’s Davis, the final speaker, repeated data released when he was appointed that Yahoo filters and doesn’t charge for roughly 12 to 15 percent of search clicks for a range of reasons. Davis was quick to point out however that click fraud or invalid clicks represent a much smaller subset of that filtering. He made appeals to the audience several times to reach out to him and discuss concerns and issues. Overall he made a very strong and credible appearance for Yahoo that the company is committed to greater openness and dialogue with advertisers and the industry about issues of click quality.
As you might expect the Q&A portion of the panel was lively, but I was surprised by the intensity of one of the audience members who was irate about “being sold a bunch of 12 bananas and only receiving six.” While his position is justified the irony here is that the tracking capabilities of search marketing themselves create expectations that don’t shadow virtually any other advertising medium. In other words, the possibility of the discovery of click fraud creates the expectation that every billed click should be valid.
One got the sense from the panel that there are no true “third parties” in this debate. Though the presentations were highly credible, the search engines have incentives to reassure advertisers that actual click fraud is extremely low to maintain confidence in their systems. And companies like Click Forensics have clear incentives to suggest the problem is growing, in part to justify their services. The two speakers who were most impressive to me were Marshall who was very thoughtful about the issues and wasn’t selling his services and Yahoo’s Davis who seemed like a man who sincerely wanted to work with and satisfy advertisers.