Click Fraud - Nuisance Clicks - PPC fraud

Click Fraud has been discussed a lot lately. In fact, it's a challenging classification problems to detect fraudulent click patterns in log-files. However, the term "click fraud" is misleading. You may go to Overture (or Google) and search for "phone cards" and click on the top listing. Did you commit a criminal act of fraud? No. The advertiser would probably have preferred that potential customers click on the listing, but nobody can force you to buy something or go beyond the first page. There is no legal mean known to me against nuisance clicks (I am not a lawyer, though!). Companies that offer Pay-Per-Click advertising try to protect against the worst kinds of abuses by carefully checking the log files of click-throughs, but that might be futile.

The German computer magazine c't reported in issue 13/2004 (page 170) about click-spam. An experiment was conducted in which their own ad was only clicked on by them in a period of two days by a script. Google did not discover this strange pattern and charged for all of the clicks.

There are many causes for PPC fraud. A study (link?) found that most users do not realize that sponsored results are being paid for by advertisers and might inadvertently click on the ads over and over again to get to the companies website. That might especially be true as people use bookmarks less and less - why bother when you can find your site again so easily?

On the other end of the spectrum we have people that try to damage their competition. There is a variety of ways of achieving this. While having humans click on ads over and over again, it is certainly possible to program to automate this, especially if the URL used to redirect users to the advertisers site (the advertising network also wants log clicks for billing) is predictable. In a recent slashdot discussion one reader posted the following comment:

   

I made a program to make fraudulent clicks for a customer who was being attacked by a fraudulent company. The company created a similar looking domain name and just ripped his entire site and put it on there.

He was going through lawyers to stop this guy but it was taking a long time. He saw that the guy was advertising on Google (right above him). So I created a program that got a list of anonymous proxies from a site, then it would randomly choose a proxy, then pick a random interval, and a random referrer (from a list), and a random browser tag. Anyway, using this program we could click on his ad-word hundreds of times a day, costing him money and making the "bad-guy" max his limit for the day.

I was actually pretty shocked that it worked, so I gave my client his money back and emailed Google about it. I talked with a few people at Google including a higher level technical guy. I described my method and he thanked me, and I've never heard from him again.

Moral of the story... none really, but I don't think these fraudulent clicks can be stopped. This is another case of a company who's business model is pretty much destroyed because of the anonymity of the Internet.

I need a new story to tell :-/

-- D3X

This is basically how the bad guys do it as well...

Thoughts...

  • PPC providers should scan for open proxies and blacklist them. Also, there are already lots of open-proxy lists available. At the very least those should be black-listed.
  • PPC providers should check if the ad was actually placed on the referring page or the originating search.
  • PPC providers should check, if a real browser clicked the ad. This could be done by executing some java-script code (many bots don't have a full blown java script engine), or doing redirects that don't occur every time. Simpler bot-scripts will be detected by this and it certainly raises the bar.
  • Advertisers should log the way of the user through the web page. Are images being downloaded? Are any other pages clicked from the initial page? Though it might be difficult to convince the advertisers to give out their log-files, it might be worth a shot if bots can be detected that way.
  • Advertisers should watch out for the latest Adwords flaw.

Links

Back to the top   [Sitemap]

This page is Copyright © Markus Breitenbach 2010. All rights reserved. Any opinions expressed here are my own and might not reflect my employers opinion.
[This page: http://cervisia.org/click_fraud.php was last modified: July 13 2006 16:13:20.]   [Home].   Email me   Visit Markus Breitenbach's other homepage.
-

ppc scam click fraud nuisance clicks