Why Google has already passed its peak

Matthew Lynn

7 October 2006

Maybe Google can afford to lose a few users. It would be a lot more worried about losing advertising revenue — the source of its huge profits. Google makes its money from the little highlighted ads you see on the side of every search result. It also sends ads to website publishers — advertisers sign up with Google, and their ads then appear on thousands of websites around the world. The advertiser is only charged when someone clicks on their ad, and the revenue is divided between Google and the website publisher.

It’s a great model. The ads are highly targeted and pretty cheap. There is one problem, however. The system is massively open to something called ‘click fraud’. Here’s how it works: you set up a cheap website, say, on gardening; Google delivers lots of ads to your website for, let’s say, garden sheds. You sit there all day pretending to be a visitor to your own website, clicking on the ads and — as the publisher — collecting several cents each time. That may sound a dull way to make a living, but there are back-street firms in India where you can pay people to click all day on Google ads; or you can design a computer program to do it for you.

Nobody knows exactly how great the problem is, because no one owns up to it. Click Forensics, an American consulting firm, reckons 14 per cent of the clicks are fraudulent; others put it as high as 30 per cent. Google, naturally, says the level is much lower. The problem is that it undermines confidence in the model: advertisers don’t want to pay for phoney clicks. The more rampant click fraud becomes, the weaker Google’s business gets.

Worse, website publishers are starting to tire of being paid per click. Newspapers and commercial television channels get paid by how many people see the ads they carry, not by how many make further inquiries. Why shouldn’t internet ads work the same way? In short, the pay-per-click model may be fundamentally flawed — but it’s the only model Google has.