Forum Moderators: martinibuster
After wining, dining, dancing and karaoke-ing at the "Google Dance" inside the search company's Mountain View, Calif., headquarters Tuesday night, 2,000 online advertising strategists came face to face with Google's Chief Executive Eric Schmidt Wednesday morning, and he wasn't nearly as much fun....he wouldn't even answer one of their most basic, burning questions: How much click fraud happens within Google's advertising system?
...Schmidt's wishy-washy claim comes at a time when his company and third-party click fraud detection firms are sparring over the prevalence and method of tracking the bogus clicks,...
With transparency being all the rage, Schmidt may be losing some credibility with the Google-loving media.
Unfortunately, the ball is in Google's court, and they want to simply sit on it.
The message is that Schmidt either won't or can't answer how much click fraud occurs. And, I've read more than one column now about Google's click fraud issues, from more than one reporter.
The message is that Schmidt either won't or can't answer how much click fraud occurs.
I'm confident that it's "won't," not "can't," and his position isn't unreasonable. If he gave out any numbers, the critics wouldn't believe those numbers, and he'd have a zillion AdWords advertisers worrying about whether their own reported numbers were above or below the average. Plus, why would he want to reveal such information to competitors?
Compare that hit piece with this MarketWatch report [marketwatch.com].
Schmidt said Google "will likely" start reporting how often its advertising customers are charged for a fraudulent ad click. This would be in addition to details Google already makes public concerning how often Google search ads are clicked on for invalid reasons..."We will eventually do the majority of what you're suggesting," Schmidt said while answering a question...
That's not answering a question, that's beating around the bush.
Is that right? Half a million? Last I heard here it was around 150,000
why would he want to reveal such information to competitors?
Which is fair. The problem is, Google wants to have it both ways. They want to a) keep secret the amount of click-fraud their data reveals and b) try to convince advertisers it's acceptably low and c) shoot down anybody who claims it's not acceptably low at all.
Kinda like Oracle forbidding their customers to publish any real-life benchmarks of the Oracle software they've purchased, but publishing a stream of their own benchmarks that show that (surprise!) Oracle software is Really Fast.
Google can try to have it all ways, but should not be surprised that they have no bully pulpit for preaching against the folks who may exaggerate the amount of click fraud seen in the field.
In related news, the idea that your click-fraud rate is a competitive secret that needs to be kept hidden should not feel reassuring to advertisers. We don't know what the true fraud rate is, but we know it ain't low enough to be a valuable selling point! :-)
Yes EFV, it is one reporter's column. A Forbes reporter.
The People Magazine of the financial world.
At least it wasn't a Dan Lyons piece, so there is a chance that it is at least partially accurate.
The message is that Schmidt either won't or can't answer how much click fraud occurs.
I'm convinced it's "can't", although he would like advertisers to believe that it doesn't matter and G can manage whatever comes up.
The truth of the matter is that fraudulent traffic can be created that looks like nonconverting traffic quite easily. Spend some time studying the Internet protocols and architecture and you will reach the same conclusion.
i would disagree with that, because the forbes article did more direct quoting from schmidt than your marketwatch piece did... in part:
""But we believe we have it under control, and we believe it is getting better. Though, we don't know what we don't know--we only know what we can detect.""
sounds pretty wishy-washy to me... how can google believe that it's getting better, when they admit that they can't even detect all of it?
>>>Is that right? Half a million? Last I heard here it was around 150,000<<<
i think those numbers are in the ballpark for ypn, per one of the ypn blogs.
If they could detect it then they could do something about it.
Therefore they simply don't know the total magnitude of click fraud because they can't detect it and can't do anything about it...yet.
I suspect
Sure, and I suspect otherwise.
Bottom line: We're all guessing.
Or did you mean to say that they know what that percentage is, given what they can detect, but simply won't release that information? That's pretty much a given eh?
Or did you mean that they can ascertain the click fraud percentage without being able to detect all click fraud?
What your statement suggests is that you 'suspect' that they can determine the click fraud percentage without being able to detect all fraud or that they can detect all fraud and therefore know the exact percentage. Which means Schmidt is prevaricating.
[edited by: digitalghost at 12:24 am (utc) on Aug. 12, 2006]
Or did you mean to say that they know what they percentage is, given what they can detect, but simply won't release that information? That's pretty much a given eh?
It isn't a given to some of the people here.
Bottom line: We're all guessing.
I'm 100% certain that they cannot algorithmically detect all click fraud.
They are possibly getting better at identifying things that may be fraudulent that a few years ago passed underneath their radar. But note that the traffic may not be fraudulent, and if they don't charge for it, they're losing money. In fact they're not only losing money on not charging, they're losing money on the cycles expended to determine if the traffic was fraudulent.
But note that the traffic may not be fraudulent, and if they don't charge for it, they're losing money. In fact they're not only losing money on not charging, they're losing money on the cycles expended to determine if the traffic was fraudulent.
Sure, but that's a cost of doing business, and they seem to be doing okay.
Yeah, we know that.
i would disagree with that, because the forbes article did more direct quoting...
You're missing my point. The point centers on what the article left out, not what they put in. As journalism goes, that piece is slanted precisely because of what they chose to leave out (not what they put in).
[edited by: martinibuster at 9:12 pm (utc) on Aug. 14, 2006]
And articles can be slanted by what the journalist decides to put in the article. So what is the point? ;)
They also make a big deal about returning invalid click dollars back to the advertiser with the funds that are withheld but what about the pay periods when the scammer was under the radar?
A big joke, my prediction is that PPC will go the way of 900#'s, regulated or outright banned.
If they want to publish REAL fraud data, show how many "publishers" have been bounced and how many dollars have been "returned" to advertisers. This whole head fake of looking at click streams is only going to work for so long.
[edited by: TypicalSurfer at 10:09 pm (utc) on Aug. 13, 2006]
And articles can be slanted by what the journalist decides to put in the article.
Good point! Like the author's use of loaded terms such as wishy-washy, which turns this article from a news piece to an editorial piece. It's important to make the distinction. In this age of Fox News, so many are so used to being spoon fed editorialized journalism that they are no longer able, or simply refuse, to discern what is editorialized reporting and what is old fashioned dispassionate reporting.
The fact that the Forbes article is bad reporting doesn't negate the opinion of anyone that happens to agree with the editorial slant. And I'm not arguing against that. I'm simply pointing out that the Forbes article is a poor piece of reporting.
If G$$G were serious about click fraud they would be filing criminal and civil complaints against fraudsters.
Hmmm ... some publishers' accounts have been terminated, and it turned out they were clicking on their own ads. Arguably, they have committed wire fraud. OTOH, the advertisers could file complaints against them (assuming their identities have been made available).
Other accounts were terminated for T&C violations, which aren't all crimes. In fact, as an example, encouraging someone to visit one's sponsor is an acknowledged means of doing business in most other ad-sponsored media.
A big joke, my prediction is that PPC will go the way of 900#'s, regulated or outright banned.
Glad you have noticed this. I am disappointed that the new click fraud "working group" isn't (yet) looking at things like this as models for characterizing the problem and recommending solutions.
Good point! Like the author's use of loaded terms such as wishy-washy, which turns this article from a news piece to an editorial piece. It's important to make the distinction.
That caught my eye, too, which is why I called it a "column." The writer's job title may be "reporter," but it wasn't a reportorial piece; it was written to support a point of view. (There's nothing inherently wrong with that, but it's important to make the distinction.)
If G$$G were serious about click fraud they would be filing criminal and civil complaints against fraudsters.
This discussion isn't about G$$G, it's about Google. Who's G$$G?
A big joke, my prediction is that PPC will go the way of 900#'s, regulated or outright banned.
I think that's unlikely, unless governments decide to regulate or ban every kind of payment and commission system that has the potential for fraud and error.
I think that's unlikely, unless governments decide to regulate or ban every kind of payment and commission system that has the potential for fraud and error.
Righto, regulation is normally reserved for those systems that have FUNDAMENTAL flaws like 900#'s and PPC. At some point gov't will step in if there is wholesale fleecing of any group (even peon adwords advertisers), theory being that fraud is "trickle down" placing burden on society and consumers, unfortunately for G$$G fanboys it is not viewed as a self correcting phenomena.
[edited by: TypicalSurfer at 3:12 am (utc) on Aug. 14, 2006]
Of course, in both cases the businesses want to minimize their losses, but successful advertisers are not going to stop advertising because of click fraud any more than Walmart is going to shut their doors becase a certain percentage of their goods walk out the door without being sold.
BigDave wrote:PPC isn't going away as long as advertisers are making money at it. Most of the ones that are using effectively just consider it a cost of doing business, just as retail stores consider shoplifting a cost of doing business.
Of course, in both cases the businesses want to minimize their losses, but successful advertisers are not going to stop advertising because of click fraud any more than Walmart is going to shut their doors becase a certain percentage of their goods walk out the door without being sold.
A percentage is assumed, like in other businesses, and anyway advertisers are not charged for invalid clicks, and reimbursed if some weren't detected in advance.
Also, the invalid click percentages seem to be low and under control, at least according to the AdWords blog (discussing a controversial third party study, etc.):
[adwords.blogspot.com...]
[adwords.blogspot.com...]
The system seems to be working quite well.
[edited by: Juan_G at 11:12 am (utc) on Aug. 14, 2006]
A small sampling of click fraud reports of the kind Google now makes available, seen by MarketWatch, suggests Google's click fraud rate is between 20% and 25%, which is in line with estimates from independent research companies.
I'm hoping this is high and a large sample would result in a lower rate.
