Due to a number of lawsuits, the topic of click fraud has recently become especially relevant. Several analytical companies immediately published the results of their own research. They estimate the level of fraud on the largest search engines at 14.1-14.6 %.
During the Lane's Gifts v. Google case, the court asked independent experts to scrutinize Google for "click fake" fraud. The assessment was conducted by New York University information systems professor Dr. Alexander Tuzhilin, who, while working on the document, visited the company's office three times and spoke with more than a dozen employees of the Click Quality team (founded in 2003). Google posted the 47-page report he prepared online (PDF file).
In his report, Tuzhilin assesses the effectiveness of Google’s anti-fraud filters as “reasonable.” He also notes that the quality of Google's anti-fraud systems is constantly improving. The qualifications of Click Quality employees are beyond doubt. Google's current set of anti-fraud filters works quite stably and requires only periodic adjustments. The system has four levels of protection with a different operating algorithm in each of them: preliminary filtering (clicks of Google employees and erroneous entries in logs), online filtering (differences from normal user behavior, non-compliance with specified rules for a normal click), automatic offline recognition system and manual offline click fraud detection system. There is no need to radically remake this system, Tuzhilin believes, even if massive fraud attacks begin on the AdWords network, as happened several times over the past year, it will cope. Moreover, it is almost impossible to radically improve it: each new filter in an online filtering system is capable of recognizing a maximum of 2-3 % of fake clicks that are not recognized by other filters.
From Tuzhilin’s report you can learn some interesting details about how Google’s anti-fraud system works. For example, an important element of it is the collection and analysis of information about the user’s actions on the advertiser’s website after he went there via an advertising link. In particular, a “key” page is marked on the advertiser’s website (for example, an order basket in a store), the visit to which is also recorded.
The report also contains criticism. According to Tuzhilin, the work of Google’s anti-fraud system cannot be called ideal. For example, as a weak point, he notes the fact that the system records user behavior only after he clicks on an advertising link. But his activity up to this point is not taken into account in any way.
In addition, Google does not show advertisers which specific link clicks it has considered normal and which are incorrect. The company's reports contain daily statistics. Recently, the AdWords system indicates the number of fake clicks and their percentage of the total, but nothing more.
Alexander Tuzhilin was sincerely bewildered by the fact that Google only in March 2005 recognized double clicks on advertising links, which were previously counted as two clicks, as incorrect. This seemingly obvious case of a fake click was considered normal in Google's system for several years, and after adding a corresponding filter, the company, by its own admission, lost a "significant portion" of advertising revenue. Such cases make us wonder: is Google really sincere in fighting Freud in its advertising network? Actually, in his report, Tuzhilin listed a number of anti-fraud methods (p. 34), which Google for some reason does not use.
During the Lane's Gifts v. Google case, the court asked independent experts to scrutinize Google for "click fake" fraud. The assessment was conducted by New York University information systems professor Dr. Alexander Tuzhilin, who, while working on the document, visited the company's office three times and spoke with more than a dozen employees of the Click Quality team (founded in 2003). Google posted the 47-page report he prepared online (PDF file).
In his report, Tuzhilin assesses the effectiveness of Google’s anti-fraud filters as “reasonable.” He also notes that the quality of Google's anti-fraud systems is constantly improving. The qualifications of Click Quality employees are beyond doubt. Google's current set of anti-fraud filters works quite stably and requires only periodic adjustments. The system has four levels of protection with a different operating algorithm in each of them: preliminary filtering (clicks of Google employees and erroneous entries in logs), online filtering (differences from normal user behavior, non-compliance with specified rules for a normal click), automatic offline recognition system and manual offline click fraud detection system. There is no need to radically remake this system, Tuzhilin believes, even if massive fraud attacks begin on the AdWords network, as happened several times over the past year, it will cope. Moreover, it is almost impossible to radically improve it: each new filter in an online filtering system is capable of recognizing a maximum of 2-3 % of fake clicks that are not recognized by other filters.
From Tuzhilin’s report you can learn some interesting details about how Google’s anti-fraud system works. For example, an important element of it is the collection and analysis of information about the user’s actions on the advertiser’s website after he went there via an advertising link. In particular, a “key” page is marked on the advertiser’s website (for example, an order basket in a store), the visit to which is also recorded.
The report also contains criticism. According to Tuzhilin, the work of Google’s anti-fraud system cannot be called ideal. For example, as a weak point, he notes the fact that the system records user behavior only after he clicks on an advertising link. But his activity up to this point is not taken into account in any way.
In addition, Google does not show advertisers which specific link clicks it has considered normal and which are incorrect. The company's reports contain daily statistics. Recently, the AdWords system indicates the number of fake clicks and their percentage of the total, but nothing more.
Alexander Tuzhilin was sincerely bewildered by the fact that Google only in March 2005 recognized double clicks on advertising links, which were previously counted as two clicks, as incorrect. This seemingly obvious case of a fake click was considered normal in Google's system for several years, and after adding a corresponding filter, the company, by its own admission, lost a "significant portion" of advertising revenue. Such cases make us wonder: is Google really sincere in fighting Freud in its advertising network? Actually, in his report, Tuzhilin listed a number of anti-fraud methods (p. 34), which Google for some reason does not use.
