Mark Cuban vs. Don Dodge on Click Fraud - round 2

Mark Cuban is at it again...another post on "Why Click Fraud is far greater than I imagined". To his credit Mark points out right up front that this is an opinion piece...no investigative reporting. Mark is a smart guy and makes logical arguments, but there is simply no proof that Click Fraud is a big problem, or "greater than I imagined".

Mark's premise is that 1) bad guys exist, 2)bad guys find click fraud easier than other scams,  3)bad guys always gravitate to where the easy money is, 4)SPAM and SPLOGS are exploding, growing every hour. Therefore, click fraud is a huge problem. Here is a quote from Mark's blog.

Now i have no idea how much money is being lost to click fraud. All i know is that when the black hats see easy money, they take it. I also know that they are greedy and a jealous bunch. The more they see the more they take, so you can pretty well bet that the amount of click fraud is going up by the minute.

And no amount of IP repetition algorithms are going to stop them.

Again, this is all opinion. No investigative reporting going on. Yet.

OK, here is my opinion. No investigative reporting from me either, although I have talked to several click fraud detection companies, and plan to dig deeper to better understand the size of the issue. If indeed it is a big problem I just might start a click fraud detection company. My first call would be to Mark Cuban to see if he is interested in funding the company.

The marketplace automatically adjusts for click fraud by lowering the price they are willing to pay for clicks. Remember, CPC ad prices are set by competitive auctions for each keyword. Advertisers run ROI's on all their ad spending. They will adjust what they are willing to pay, up or down, based on the real returns from each campaign.

Google has incentive to stop click fraud. A common "logical" argument of the click fraud alarmists is that Google has no incentive to stop click fraud. Well, in fact they do. I just explained the incentive in my previous point. Click fraud will cause the price per click to drop, which will lower Google's revenue. Google wants higher prices for quality clicks, so they have every incentive to stop click fraud and keep prices higher.

Greed always exposes click fraudsters. Mark Cuban makes logical arguments about how click fraud can be done. I agree it can be done without detection on a small scale. But, greed causes the bad guys to crank up the frequency, which causes them to be caught by the detection algorithms. Google's algorithms monitor everything, and can detect abnormal traffic or patterns that are likely click fraud. There are lots of stories about ad network sites being banned from Google due to suspected click fraud. Google, and other CPC ad networks, take click fraud very seriously and detect most of it before it ever gets billed to their advertisers.

Click fraud is far lower on quality sites. Major sites like Google, Yahoo, MSN, AOL, and others have very sophisticated monitoring tools that analyze everything.  They can easily detect click fraud and stop most of it. Tier 2 sites are also pretty good at it. There are of course a class of web sites that are set up solely for advertising fraud. Anyone who allows their ads to be run on these sites gets what they deserve.

SPAM and SPLOGS stink, but it isn't click fraud. Everyone agrees that SPAM stinks and that SPAM has invaded blogs creating what is called SPLOGS, for Spam Blogs. However, linking SPAM to click fraud is apples and oranges.  They are simply not the same.

You might recall that Mark Cuban and I got into a debate on this subject a couple months ago. Here is my original post (see Mark's comments to the post) and Mark's post on his blog "Blog Maverick". At the end of round one I suggested that this debate should be settled by jump shots from the top of the key. Mark owns the Dallas Mavericks and is a huge basketball fan. So am I. His Mavericks just won the Western Conference Championship and are going to the NBA Finals for the first time in the teams 26 year history. Congratulations to Mark and the Mavericks.

I still think I could whoop Mark in a 3 point shooting contest and settle this debate once and for all :-)

Subscribe - To get an automatic feed of all future posts subscribe here, or to receive them via email go here and enter your email address in the box in the right column.

Comments

[underscorebleach]

"Remember, CPC ad prices are set by competitive auctions for each keyword. Advertisers run ROI's on all their ad spending."

- Wrong. Many AdWords advertisers run ads and pray. You don't know how many actively track ROI or have a clue how to monitor their ad buys' effectiveness.

"Click fraud will cause the price per click to drop, which will lower Google's revenue. Google wants higher prices for quality clicks, so they have every incentive to stop click fraud and keep prices higher."

- Really? You're telling me that they'd rather have 100 legit clicks than 100 legit clicks + 20 illegitimate clicks? That's gonna hurt them?!

"Greed always exposes click fraudsters."

- This *is* an opinion piece, isn't it? If I were running a click fraud scheme, I'd hook up with the virus makers in the Russian Mafia and get zombie PCs to click. Too bad the Google guys will absolutely, positively catch me. http://tinyurl.com/pstm6

"Click fraud is far lower on quality sites."

- Mmm. Well, that's a good thing, because Google doesn't make any money from (for example) mistyped domain names. http://tinyurl.com/pea7m

"SPAM and SPLOGS stink, but it isn't click fraud."

- O RLY? I was pretty sure that splogs were set up to gain in the SERPs, get traffic, and either monetize that traffic through native ads or by pointing to other sites with ads.

But I guess I just don't know what I'm talking about.

[TheShadow]

Don - so you haven't noticed how spam has mushroomed out of control?

Why is that?

Could it be - that spam actually makes someone money?

Hmmm...and you think those spam masters are legitimate businesses, right??? Tell me you don't believe that Don!

Are you saying click fraud is not significant?

Do you honestly think that there are not significant numbers of bad guys out there scamming the system to the tune of millions of dollars?

Seriously - I want some of want you are smoking.

When I read about some peasant in a village in India who set up a web site - and is making over $1,500 USD per month on google ads that are clicked on his web site - I have to ask WTF?! - and apparently that isn't an isoldated instance...

Don - how many times have you actually clicked on one of those adsense ads? Me? probably once or twice in the last 3 years.

So where are all those clicks coming from?

[Don Dodge]

Spam, email or otherwise, is not click fraud. Spam requires a user to take action and click on a URL or an ad. When a human user takes action it isn't click fraud. It may be stupid, but it isn't click fraud.

Yes, Spam is increasing and it has invaded blogs. The cost of generating the Spam is very low so it doesn't take much of a response rate to generate a little profit. Personally I have never knowingly clicked on a Spam URL or ad. They are just ridiculous. I can't imagine how any person intelliegnt enough to use a computer would click on these dumb Spam ads.

The marketplace ultimately corrects all of these issues. Users get smarter and don't click. Advertisers lower their bids for words to the poinnt where it isn't profitable for the click fraudsters. Google and the Ad Networks get better at detecting and eliminating click fraud so as to preserve higher ad rates.

Yes, Click Fraud exists and can go undetected in small amounts. The stories of how easy it is to perform click fraud are theoretically true. However, the ideas don't scale because that is when they trip the detection algorithms and get caught.

Click Fraud has become an urban legend born on the Internet, much like the the welfare mother in Chicago who was collecting 20 welfare checks for children she didn't have. Ronald Reagan told the story eloquently hundreds of times on the campaign trail. The only problem is that it wasn't true. When smart people like Mark Cuban and technical people in the industry make persuasive arguments about how Click Fraud works...people automatically believe them. Where is the proof? Are there isolated incidents? Sure. But where are all these rich CF entrepreneurs? They are merely urban legend.

[Dexter Johnson]

The biggest question is, if there is little or no click fraud, who in the world clicks on these ads? Seems as if noone knows a single person who has clicked on a google ad more than once in the past year.

[Don Dodge]

Millions of people click on web site advertisements, and the vast majority of them are real interested consumers. This activity generates billions in revenue for Google, Yahoo, Microsoft, AOL, and others.

We are talking about obvious SPAM ads and sites that are nothing more than a page of links and ads wrapped around some trivial content. I have no idea who clicks on such ads, and I doubt that many people do.

[Rick]

I'm an equity analyst that has to constantly worry about/value/price clickfraud issues. Being in the investment world, I'm a big believer that advertisers can price protect (i.e. if clickfraud is solved, keyword prices will simply go up). But really, even the stock market isn't efficient in the short run. Furthermore, in a growing market, even advertisers that are shifting dollars away from online (e.g. FTD and Blue Nile) are easily replaced by new advertisers. The real way to measure impact of clickfraud would be to measure advertiser churn.

An even bigger issue is, have clickfraud issues just begun? In the constant battle between clickfraudsters and ad platforms, do the clickfraudsters have more to gain?

[TheShadow]

Perhaps you would care to explain away the analysis of this company's efforts?

http://www.clickfraudindex.com/

[TheShadow]

Don,

Are you saying that radiator.com is lying?

http://multichannelmerchant.com/webchannel/seo/5-17-06-Radiator/

[TheShadow]

You want evidence?

How about this?

http://www.pbs.org/cringely/pulpit/pulpit20060525.html

[Don Dodge]

Mr. Shadow, do you have a point you would like to make? I don't see any discussion points in your numerous "comments". I followed the links in your "comments"...nothing new there.

Of course Radiator.com is not lying. But they don't have indisputable proof of every instance of click fraud either. They can only suggest what might be click fraud. Some of it will be obvious to both parties. Other instances will be debatable. I suspect that Google will investigate the issues and resolve them with Radiator.com. There was no resolution at press time of that article, but sometimes things don't resolve as fast as customers would like. In the end Radiator.com can decide to take their advertising budget to Microsoft or Yahoo if they are not happy with Google. Being from Microsoft I hope they bring their business here.

Yes, click fraud does exist. No argument there. The debate is about the extent of click fraud. Cuban thinks it is huge, and I don't. I think the marketplace (advertisers, users, ad networks) naturally resolves these issues.

Advertisers have all the power. They are the ones deciding how much to spend and who to spend it with. There are competitive ad networks out there. Microsoft has announced its own ad network to compete with Google. I would highly recommend any advertiser who is unhappy with Google should try Microsoft or Yahoo.

[CPCcurmudgeon]

I think you have underestimated the extent of click fraud. It has been a problem for many years; before the dotcom boom, even. It is just getting press now because of lawsuits.

Something for you to think about: If you started an online ad campaign and saw a gradual increase in impressions and clicks, how would you determine whether or not it was a result of fraud or increased popularity?

My general opinion is that charging per click, impression, or other discrete event is the real culprit here. Charge fixed fees and it all goes away. (Pay per call is vulnerable to fraud perpetrated by people who are willing to call for a few cents a day, not unlike people who are willing to click for a few cents a day. Furthermore, the SEs/ad networks lose money on those payments that can't be tied to a call, e.g. someone hears about the company from a friend.)

[underscorebleach]

"When smart people like Mark Cuban and technical people in the industry make persuasive arguments about how Click Fraud works...people automatically believe them. Where is the proof? Are there isolated incidents? Sure. But where are all these rich CF entrepreneurs? They are merely urban legend."

Don, there's a big problem with your argument. You ask "Where is the proof." The proof is locked in shadowy, mysterious lair of Google's database. They're the most secretive company I know -- do you know what kind of NDA's they make job applicants sign?!?!? -- and you're willing to take them on their word, eh?

"The marketplace ultimately corrects all of these issues."

Same happened with the DotCom Bubble. Look for the same to happen with text ads. The success of AdWords/AdSense has been driven by naive people who want to "stick an ad on the Internet" -- not by people who do rigorous ROI studies on their ads.

[CPCcurmudgeon]

A couple more points I didn't have a chance to make before:

I don't think comparing click fraud to waste in traditional ad
spending is a good analogy. For one thing, let's assume that click
fraud went away completely. There would still be waste in CPC
advertising -- people who click on sites but don't buy anything for
legitimate reasons (costs too much, wasn't what they wanted, etc.).
Likewise for CPM advertising -- people who ignore the ads or filter
them out somehow. Another issue is that in radio and TV advertising,
the ad spend is based on a statistical projection of the audience
size. No individual viewer or listener can markedly influence the ad
spend. In fact, it isn't even in a viewer or listener's best
interests to game Arbitron or Nielsen rankings by somehow lying about
what they listen to or watch, because they lose the ability to
(statistically) keep the programs they want on the air. However, the
same is not true for CPC, CPM, even cost-per-action (a non-sale action
such as a registration). It is relatively easy to generate automated
clicks, either by people who are working in conjunction with each
other, or using compromised computers. Those clicks can mimic normal
(nonconverting) behavior. It isn't even necessary to use automation;
a group of individuals topologically dispersed (from an IP address
perspective) can manually do the same, especially for high-paying
clicks.

Regarding the marketplace sorting this out, sure, perhaps someday it
will. But does that mean advertisers who were defrauded shouldn't get
refunds? Keep in mind that the lawsuits filed go back to a time
before click fraud was widely discussed. Some advertisers were given
refunds, but others were not, and it was at the search engine's or ad
network's discretion who was refunded. Plenty of people gave what
they considered to be reasonable proof that they were overcharged, but
they didn't get refunds, or didn't get as much as they thought they
were owed. So at the very least there should be (IMO) an
investigation into what criteria were used to issue refunds and
whether it was fairly applied to all advertisers.