Click Fraud isn`t Going Away - First, Admit There's a Problem
(Page 4 of 4 )
According to Richard, “Google, Yahoo! and MSN are stonewalling on click fraud, to their own and others’ detriments.” In short, the companies need to admit what the respondents to the Outsell survey already know: click fraud is a problem, it’s a big problem, and, despite their protestations to the contrary, the search engines are not being effective enough in their efforts to fight it.
But that’s just the first step. “The industry must openly admit that there is a major problem, adopt independent audits and open itself to new approaches created through collaborative solutions with [Google, Yahoo, and MSN], publishers, advertisers and the hacker-geek community.” In other words, it’s not something each individual search engine can hope to successfully tackle by itself. Fortunately, recent news seems to indicate that the major search engines have at least become more open to collaborating with each other to take the first steps toward a solution.
Think about the spam problem, for example. We’ve been fighting that for much longer, and it still hasn’t gone away, perhaps for some of the same reasons. “The fraudster community is like the hacker community in that it always seems to stay one step ahead of the technological traps thrown in its path…As long as you can get paid for generating a click or harm your competitor with a click that has no value, the pay-per-click system will be gamed.”
That makes it all the more interesting that major search engines are looking at different models. Google, for instance, is testing a pay-per-action system, where publishers do not get paid unless the surfer who clicks through the ad on their site actually does something – leaves information, signs up for a newsletter, buys a product, what have you. There seemed to be relatively little interest in this model six months ago. Now, however, with two click fraud settlements made, I believe we will see more interest in pay-per-action as a way to deal with click fraud -- maybe even once and for all.