Google says filters reduce click fraud to 1 in 5,000 clicks

Just 0.02% of clicks that Google charges to advertisers are the result of fraud, the company has said. The figure is a fiftieth of one per cent despite the fact that click fraud attempts constitute just under 10% of all ad clicks.

Text adverts displayed next to the results of searches typed into Google's search engine are the web giant's main source of revenue. A problem with the adverts is the way they are billed. A company is charged each time someone clicks on its ad to visit its website.

While this makes for a very accountable way of advertising, where the company pays only for those people interested in its products or services, it is open to fraud.

A competitor could employ someone or write software to continually click on an advert in order to force the advertiser to pay each time. This has been used as a way to run down a competitor's advertising budget.

Another form of click fraud focuses on AdSense, the Google product that lets website publishers display Google ads on their own sites. When visitors click the ads, Google shares the ad revenue with the publisher – which is an incentive to unscrupulous publishers to click their own ads.

Click fraud, which can affect any providers of pay-per-click (PPC) advertising, has been the subject of lawsuits and has long been viewed as a threat to click-through advertising if not fully combated. As operators of the largest online advertising networks, Google and Yahoo! have faced the most criticism.

Google has for some time published the figure that just under 10% of clicks are invalid, and could be fraudulent, which represents up to $1 billion a year in advertising revenue.

What the company has never before revealed, though, is how many of those invalid, or potentially fraudulent, clicks are filtered out by its system. It now says that it stops all but 0.02% from getting through its system, using filters and off-line analysis. That means that only two out of every 10,000 advert clicks that are charged for are fraudulent clicks.

"Our click quality team investigates every inquiry we receive from advertisers who believe they may have been affected by undetected click fraud," said Shuman Ghosemajumder, Google's product manager for trust and safety, on the official blog of click advertising product AdWords.

"Many of these cases are misunderstandings, but in most cases where malicious activity is found, the clicks have already been filtered out and not charged for by our real-time filters," said Ghosemajumder. "Because of the broad operation of our proactive detection, the relatively rare cases we find of advertisers being affected by undetected click fraud constitute less than 0.02% of all clicks."