An Arkansas court is slated to begin a two-day hearing today of the settlement proposed by Google in a class-action click-fraud suit first brought by Lane’s Gifts and Collectibles of Texarkana AR.
Under the proposal, Google would set aside $90 million to settle all claims of click fraud by its AdWords pay-per-click advertisers dating back to 2002. Advertisers whose complaints were upheld would get credits that they could apply to future Google search advertising, rather than cash. About $30 million of the settlement would go to pay legal fees and court costs.
Some Google advertisers have objected to the size and nature of the settlement, and their lawyers are expected to be heard during the “fairness hearing”. One group of about 40 advertisers, represented in part by the Atlanta firm of Chitwood Harley Haynes, has complained that the settlement is based on faulty calculations of the size of the fraudulent click problem at Google. They also charge that the notices Google sent out about the proposed settlement looked like spam, so that many went unopened, leaving potential claimants uninformed about their rights in the class action.
Over the weekend, Google’s official blog posted excerpts from an independent study on the company’s fraud detection mechanisms. Commissioned as part of the Lane’s settlement, the report by New York University professor of information systems Alexander Tuzhlin found that the search engine’s efforts to combat click fraud are “reasonable” and wrote that he could say, after conversations with the Google fraud prevention team, “with a moderate degree of certainty” that the issue is “under control”.
Tuzhlin also applauded Google for policy changes back in March 2005 that ended double-billing to advertisers for customers who clicked twice on a search ad. While he criticizes the company for taking two years to make the changes, “despite its noticeable negative effects on its financial performance, Google decided…not to charge advertisers for the second click, which was an appropriate action to take,” he wrote.
However, Tuzhlin also pointed out that the cost-per-click business model is “inherently vulnerable to click fraud” because there’s no practical way to define an invalid click. Even if there were, he says, the search engines wouldn’t be able to share that definition with advertisers for fear that fraudsters would learn how to commit frauds more effectively.
