Post I.T. - Washington Post Technology Blog Frank Ahrens Sara Goo Sam Diaz Mike Musgrove Alan Sipress Yuki Noguchi Post I.T.
Tech Podcast
The Bloggers
Subscribe to this Blog

Google reveals a little more on click fraud

Sara Goo

Google has always claimed that click fraud is not a problem, at least for Google. Yet it never really provided any numbers indicating the volume of click fraud the company detected or how much money Google refunded to advertisers because of the problem. So it was anyone's guess as to how big the problem was.

Maybe I should back up here. What is click fraud and why should you care? Google makes money every time people click on the "sponsored links" places next to search results, which are really text ads paid for by advertisers. These ads appear not only next to search results but also as text ads that Google places on blogs or other Web sites, labeled "Ads by Google." (I never click on these ads but apparently, a lot of people do because it provides Google with a significant chuck of its revenue.) Each time someone clicks on the ads, the advertiser pays Google and the Web site displaying the ad. But fraud occurs when people repeatedly click on ads with no intention of really viewing the ad; rather, they want to drive up the cost for the advertiser and increase revenue for the Web site owner. It's more sophisticated than it seems. People create sophisticated "click robots" to make it look like authetic clicking when it's really not. People have also created networks of clickers, or people paid to click on ads to evade Google's click fraud filters.

It's an industry-wide problem--not just for Google--and one that has been compared to the days before newspapers and television had any accountability to advertisers to reveal true numbres of how many readers or viewers they were delivering to advertisers. Similarly for the online advertisers, there is no independent organization now monitoring how many clicks are fraudulent and many advertisers don't have the technology to know the difference.

The big Internet firms --Google, Yahoo, Microsoft and agreed to form a working group and come up with some standards to define click fraud through the Interactive Advertising Bureau.

Today, Google came a little bit closer to revealing its click fraud numbers. Shuman Ghosemajumder, who leads Google's anti-click fraud effort, said Google is consistently identifying less than 10 percent of all clicks as click fraud and it does not charge advertisers for these clicks. In fact, advertisers never know about these clicks. In addition to that, he said Google has investigated and refunded $2 million last year to advertisers who request that the company investigate certain suspicious clicks, accounting for .02 percent of all clicks.

Google said it supports the idea of creating an independent scientific ratings process with the indepedent Media Ratings Council that would create an equivalent of a Nielsen for the online search advertising firms. "There are a lot of folks who probably think we would not submit our systems to that kind of an audit and we would," Ghosemajumder said. "It's not only important for us to do a good job of managing this issue for advertisers, but to provide as much transparency as we possibly can."

OK, but what about the others? Now it's time for Yahoo, Microsoft and to fess up on their numbers.

By Sara Goo  |  March 1, 2007; 6:54 AM ET  | Category:  Sara Goo
Previous: Steve Case 2.0 | Next: Dispatches from the IP Wars

Add Post I.T. to Your Site
Stay on top of the latest Post I.T. news! This easy-to-use widget is simple to add to your own Web site and will update every time there's a new installment of Post I.T.
Get This Widget >>


Please email us to report offensive comments.

10% can be a significant discrepency, especially if we think on the grander scale of things. However, discrepancies have always existed in the advertising world. I remember the early days of banner advertising and the errors that were being reported based on the use of inadequate ad-serving technologies, inability to count through network and browser cache, etc. The fact is that as technology evolves we can expect to gain more relevancy and accuracy with online advertising. Until then, I say let 10% ride. And, I applaud Google and company for taking a proactive stance and refunding their customers.

Posted by: Kam | March 1, 2007 8:14 AM

Google still needs to create a standard where they are actually creating a process that allows all advertisers to properly reconcile the data that is being submitted to Google as inaccurate. A perfect example of this would be to have a side by side reconciliation for any given month of service in question. The process would allow both Google and the advertiser to properly decipher which clicks were billed and which clicks simply still need to be refunded outside of what Google is reporting. Google still remains biased to how it reports and filters Click Fraud, so having a 3rd party Click Fraud report to resolve any discrepancy would allow Google to understand more about their affiliates that are failing the advertisers as well as adjusting the advertisers bill for additional Click Fraud detected.

Click Defense

Posted by: Click Defense | March 1, 2007 8:37 AM

I find it a bit ironic that there is google ads at the bottom of this article :)

Posted by: Stephen Kelly | March 1, 2007 9:18 AM

This is no such thing as a 'click' - the triggering of a hyperlinked redirection can be done automatically through scripting or customized browsers without human intervention and there is no real way to tell whether a human or computer triggered the hyperlink redirection. Maybe I need to create a YouTube video to demonstrate this.

Because of this flaw in pay-per-click advertising, smart computer programmers can simulate people "clicking" on these ads and there is no real way to tell, unless you are standing over that exact computer performing the "clicking"

And last - no one want to even consider that a pay-per-click provider can basically create their own astronomical revenue generation quarter after quarter based on this flawed model, but highly hackable advertisement model.

Posted by: Ed Dunn | March 1, 2007 10:26 AM

googles principal revenue stream comes from advertising or advertisers. microsoft and yahoo profits are less dependent on clicking, advertising, and advertising, therefore fraud may not be a problem.

Posted by: egalitaire | March 1, 2007 2:59 PM

Yeah, right. Of clicks that were invalid and the people that recognized them, only .02% of those people were able to convince google to give them any money back. Isn't that what this is *really* saying?

Posted by: Jason P. | March 1, 2007 3:50 PM

forget the click through fraud issue -- how about a story on how google's beloved gmail has been down all freakin day here in the mid atlantic! grrrr....

Posted by: Tim | March 1, 2007 4:56 PM

Am I naive to think that the source ip address cannot be associated with each and every click?

Every advertiser should be assured that their click count includes at max only one click/day/terminal ip address. Multiple clicks from the same terminal over any length of time should be obvious for accurate click counting and prosecuting fraud.

Posted by: Bob Payne | March 1, 2007 5:04 PM

imagine a company that earns revenue pay per click. this company could share revenues with employees or is publicly traded. neither the employees nor the sharholders would have an incentive to click on twenty or so websites every time they went home, now would they?

Posted by: Anonymous | March 1, 2007 10:36 PM

agree w/ bob paynes comment ref ip@ etc.. imo the fundamental thing is that there is no service agreement between google et al and the real end-user.therefore there is no basis for fraud. it is reasonable that if a user uses a search engine he should adhere to R&R incl. absence of fraudulent intentions in exchange for using these wonderful search facilities.

Posted by: eric visser | March 2, 2007 7:55 AM

agree w/ bob paynes comment ref ip@ etc.. imo the fundamental thing is that there is no service agreement between google et al and the real end-user.therefore there is no basis for fraud. it is reasonable that if a user uses a search engine he should adhere to R&R incl. absence of fraudulent intentions in exchange for using these wonderful search facilities.

Posted by: eric visser | March 2, 2007 7:57 AM

If they are going to want legislation to deal with this problem I suggest tying opt-in anti-spamming legislation to this. I'm rather sick of the advertising industry getting a free ride at the expense of everyone else because they managed to engage in K street bribery.

Posted by: Gentry | March 5, 2007 3:21 PM

Discerning IP won't work. PPC providers claim to do this already and there is a very easy work around - just get a new ip address before each session. People also don't realize there are more advance autosurf syndicate setups such with VPN to grab IP from other computers worldwide using several dial-up ISPs to get ip address to simulate users from all over the nation.

It can take one person from Romania to VPN or dial into US computers to simulate a USA person clicking on ads per hour. And when you charging .50/click and sharing with 'affiliates' - do the math.....

It will be certain pay-per-click advertising will be exposed as the biggest fraud system out there. It will be also understood why it was not exposed by mainstream media after noticing the mainstream media embrace these pay-per-click ad system for revenue.

Posted by: Ed Dunn | March 6, 2007 8:44 AM

if shareholders or employees benefit just clicking just once, it is fraud. be careful not to click on the same link twice during one session and just click on as many different links as you can during a single session. foxes like to guard the chicken house. 100 employees * 100 links * 365 sessions * three cents a click. this one is even better: 10,000 shareholders and their families * 1000 links * 365 sessions * 2 cents a click -- its the perfect business model but i wouldnt put my advertising money there!

Posted by: Anonymous | March 6, 2007 3:17 PM

i didnt make it clear: clicking once in order to profit is fraud even if it is just once on each link on each day and your IP address shows up innocently enough. if there are enough people involved, the numbers are astounding -- do the math.

Posted by: Anonymous | March 6, 2007 3:38 PM

what is proposed seems only possible, and there is no knowledge of or no belief that anyone is committing fraud. the point is that pay-per-click advertising appears easy to abuse. appreciate the comments as meaningful topics for discourse and contemplation.

Posted by: Anonymous | March 6, 2007 4:00 PM

Thank you for such nice site!
chesapeake bay golf course
chesapeake bay golf
[URL=]chesapeake bay golf course[/URL]
See you!

Posted by: [*data/names.txt*] | March 14, 2007 11:20 AM

A source IP address isn't sufficient to prevent click fraud. HTTP has an X-Forwarded-For (XFF) request header which is used to identify requests that originate behind proxies. Although I have yet to verify this first hand, I've read that manipulating the XFF header permits simple source IP spoofing without the hassles of VPNs.

Posted by: Emerson Farrugia | April 2, 2007 7:52 AM

The comments to this entry are closed.


© 2010 The Washington Post Company