March 10, 2006
Click fraud

Do you know what click fraud is? It just cost Google ninety million bucks.

Google has agreed to pay $90 million to settle a lawsuit alleging the online search engine leader overcharged thousands of advertisers who paid for bogus sales referrals generated through a ruse known as "click fraud."

The proposed settlement, announced by the company Wednesday, would apply to all advertisers in Google's network during the past four years. Any Web site showing improper charges dating back to 2002 will be eligible for an account credit that could be used toward future ads distributed by Google.


Mountain View, Calif.-based Google makes virtually all of its money from text-based advertising links that trigger commissions each time they are clicked on. Besides enriching Google, the system has been a boon for advertisers, whose sales have been boosted by an increased traffic from prospective buyers.

But sometimes mischief makers and scam artists repeatedly click on specific advertising links even though they have no intentions of buying anything. The motives for the malicious activity known as click fraud vary widely, but the net effect is the same: advertisers end up paying for fruitless Web traffic.

The lawsuit alleged Google had conspired with its advertising partners to conceal the magnitude of click fraud to avoid making refunds.

The frequency of click fraud hasn't been quantified, causing some stock market analysts to worry Google's profits will falter if it turns out to be a huge problem.

Google executives have repeatedly said the level of click fraud on its ad network is minuscule — a contention that the proposed settlement amount seems to support.

The $90 million translates into less than 1 percent of Google's $11.2 billion in revenue during the past four years.

Google's had a few glitches lately, hasn't it? This article from the Sunday SF Chronicle, which I read while I was out in CA last week, has more on the state of the business in Mountain View. Check it out.

Posted by Charles Kuffner on March 10, 2006 to Technology, science, and math | TrackBack

I'm having trouble wrapping my head around this. At first blush it sounds like a pretty frivolous suit.

What's the distinction between someone who clicks on a link, then upon taking a closer look, decides not to buy (something I personally do rather frequently), and someone who clicks on a link because they dislike the sponsor and wants to cost them a penny or two? (Sort of like returning empty business reply envelopes from GOP candidates. Hmm.... maybe we should all search for "Republican" and click through every sponsored link just for the hell of it. And do so with a search engine we don't like, in case the GOP decides to sue for "click fraud.")

There are a few obvious safeguards: not counting multiple "clicks" from the same IP address within, say, a 10-second period, or non-persistent cookies with a 10-minute timeout. But beyond such obvious things, how else is Google - or anyone who gets paid on a "per click" basis - supposed to assess the "intent" of each, uh, clicker?

I would think anyone willing to pay "per click" would have reason to know that a few clicks are going to be ill-intentioned, and take that into account when negotiating a pay rate. I don't think we should assume Google failed to exercise due diligence just because they settled rather than going to court.

Posted by: Mathwiz on March 10, 2006 3:04 PM