AdWords advertisers and AdSense publishers are both greatly aware of the problem of click-fraud and invalid clicks. We can add those who are perpetrating the invalid clicks, though of course for different reasons as their concern is more along the lines of how to get away with it. Click-fraud has made it to the mainstream as well with media coverage of the recent lawsuit by advertisers against Google. So how is Google doing when it comes to detecting invalid clicks?
Google reported Friday on the Google blog that as part of it’s recent $90 million click-fraud settlement they agreed to have an independent evaluation of it’s methods and procedures for determining invalid clicks. Dr. Alexander Tuzchilin, Professor of Information Systems at the Stern School of Business at NYU was the independent expert and you can find his report about whether of not Google’s efforts to fight click-fraud in this PDF
I just read through the report, all 47 pages of it and it makes for some interesting reading. At least if you like this sort of stuff like me. Maybe it’s just my scientific background from years as a Physics and Engineering student. Or maybe I’m just a little weird. Either way or simply because of my interest in search engine advertising I found the paper a worthwhile read. The report is a little dry as you would expect of any paper of this sort, but not too technical and easily understandable.
It starts with a little history of the internet in general and then more specifically a brief history of search engines and Google, followed by the AdWords and AdSense programs. It discusses the problem of invalid clicks for both programs each of which have different motivations for making fraudulent clicks. The report continues by discussing the difficulty of detecting or even defining invalid clicks and then some general concepts about what Google is doing to combat fraudulent clicks.
Google can collect a lot of information about the clicks, whos doing the clicking, and what happens after the click, but they can’t know everything. Even more a certain amount of intent by the clicker needs to be known in order to determine the validity of any given click.
The ultimate findings of Dr Tuzhilin are that what Google is doing to fight the problem is reasonable though there are some places where he believes they could be doing a little more. While not perfect, given the difficulty of defining the problem and the difficulty of ever being able to determine 100% of invalid clicks, Google is doing a reasonable amount to prevent click-fraud.
If you’re interested at all in the subject of click-fraud you might want to check out the PDF even if only to skim it for the details that concern you most.