Google’s clicking time bomb
Google isn’t an overt, book-cooking scam like Enron, WorldCom or Computer Associates. But it is the victim/beneficiary of the fact that a meaningful, yet unknown, portion of its $9.2 billion in sales over the last 12 months is based on phony clicks on its keyword ads.
Click fraud, as the phenomenon is known, is committed by bad guys around the world who build web sites; sign up to display Google’s ubiquitous text ads, and then create computer programs to relentlessly click the ads, turning thousands of nickel-and-dime transactions into wads of serious folding money.
Advertisers know this is happening. Investors know this is happening. And Google – as well as competing services operated by Yahoo, Microsoft and others – most certainly know this is happening. But no one, except possibly Google and its confreres, knows how much it is happening. And if they know, they’re not telling.
The Economist magazine estimates click fraud at 10% to 50% of all online ad transactions. With the range so broad as to be useless, we are forced to resort to anecdotal data to get a fix on the problem. In that vein, here's the case of a friend who operates a small web site in Northern California:
My friend recently was surprised to get an email from Google urging him to increase the limit on his monthly credit card authorization, so he could pay for more clicks on the AdWords be buys to drive traffic to his site.
Intrigued by the sudden popularity of his ads, he quickly learned from Google that about $150 in ads had been clicked in the last 10 days, as compared with average bills of $23 in each of the last three months. In other words, his average daily ad bill increased almost twentyfold in the 10-day period to $15 from the ordinary 77 cents.
After ascertaining that no spike of activity on his web site matched the surge in his bill, he commenced what has turned out to be an extended tussle with Google to get a refund on the clicks that he believes are fraudulent. Google has said in a series of apparently canned emails that its proprietary algorithms show the clicks came from dozens of random web sites and, accordingly, are valid.
My friend theorizes that Google’s algorithms were fooled by cyber-bandits operating phalanxes of computers that can vary and mask the origin of clicks, so as to give the appearance that the deluge spontaneously emanated from hundreds of unrelated and legitimate users.
I think he is right. If you do, too, then his experience suggests that Google’s stock is overvalued. Here’s why:
Assuming the activity on my friend’s ads is normal over 12 months, he would pay Google $280 a year. If Google forces him to pay the additional 10-day bill of $150, then his total advertising outlay for the year would be $430, making his annual ad bill almost 35% greater than his ordinary expenditures suggest it should be.
If something like this happened to every Google advertiser at some point during the year, then logic suggests that a third of the company’s ad sales are bogus. Of course, there is no known evidence to support such a conclusion and I’m in no position to say this indeed is the case.
Still, the click-fraud issue could play havoc with Google's stock. If enough investors got it into their heads one day that, say, 10% of the clicks were phony, the stock would drop $50 a share, based on the metrics that now support its $144 billion value. Once the skeptics triggered a sell-off, there’s no telling how ugly things might get.
In considering Google’s predicament, you find yourself appreciating the safeguards designed to reassure the advertisers who patronize the legacy media.
Imperfect though the systems may be, industry-funded agencies independently measure the audiences of newspapers, magazines, television and radio. Print ads are validated with tear sheets, while program logs and air checks serve the same function for broadcasters.
Google should put a priority on doing the same for its ads. Until then, every bogus click could be the tick of a time bomb.