Last week, Google made a startling claim, one that begs explanation even after close examination. Faced with lingering questions about click fraud, Google released a blog post stating actual click fraud charges are ridiculously lower than those projected by third party analysts from the search marketing community.

According to Google, actual click fraud only accounts for 0.02% of all click activity found when Google's team is asked to audit an advertiser's account. If correct, Google's 0.02% assertion places the perceived dollar value of actual click fraud in their system somewhere in the range of $2,100,000.

In a post to the Inside AdWords blog, "Invalid Clicks - Google's Overall Numbers", Google gave a brief outline of how it claims AdWords' 100+ data-point click detection system works to filter out 99.98% of the nearly 10% of overall clicks Google determines to be invalid.

Other industry analysts have pegged the number between 15% and 20% with the click fraud index maintained by Click Forensics suggesting, "The average click fraud rate of Pay Per Click advertisements appearing on search engine content networks was 19.2 percent for Q4."

Google has never given specific figures citing how much of the PPC-traffic they filter is caught and disregarded as invalid click activity saying it fluctuates from month to month but is below 10%. Though evidence of these clicks might appear in server logs or be noted as click fraud by third party analysts, Google says it does not charge for them. In the absence of a fee charged to an advertiser, the most such invalid click activity can be called is attempted fraud.

On one side, a large third party verification system suggests a very high percentage of click fraud. On the other, Google says that invalid clicks represent less than 10% of all AdWords traffic and of those, only 0.02% get past Google's detection filters.

The true numbers likely sit somewhere in between but when it comes to understanding click fraud by the numbers, Google tends to play the numbers down. Though Google is the world's largest repository of information and a publicly traded company running the largest advertising system online, it is amazingly secretive about its data. Google has gone as far as balking at pursuing any legal action against cyber-criminals if such action might expose their systems in open court.

Proving or disproving click fraud is next to impossible without access to the data Google holds closest. Unfortunately, the only numbers we have to work with are the ones Google gives us.

Google Product Manager for Trust and Safety, Shuman Ghosemajumder uses Google's current revenue rate when he notes that, "...every percentage point of invalid clicks we throw out represents over $100 million/year in potential revenue foregone."

This is an important point to remember as we slow down the spin to ask some critical questions about the 0.02% claim. First, let's cover some things we do know.

While Google can claim to have cut the rate of actual click fraud to 0.02% of all clicks, it is unwilling or unable to provide hard evidence to prove the point. Instead, they provide images.

According to Google's year end financial statements, Google saw revenues of $10.6 billion. Of that, $10.5 billion was generated through AdWords advertising. Assuming that the rate of known invalid clicks is about 10%, Google expunges about $1.05 billion in suspicious click activity each year. 1% of Google's PPC activity equals approximately $105 million, similar to the number Shuman Ghosemajumder cites.

Google puts a great deal of effort into detecting and deleting pay per click charges stemming from invalid clicks. We know that they recognized the enormity of the problem early on. In December 2004, chief financial officer George Reyes told a Credit Suisse First Boston investor conference that click fraud could critically damage the PPC model.

CNN quoted Reyes saying, "I think something has to be done about this really, really quickly, because I think, potentially, it threatens our business model."

According to the article, Reyes went on to say, "There's a lot of bad guys out there that are trying to take advantage of this and it costs, I'm sure not just us, but eBay and Yahoo! and Amazon and the whole crowd, you know, tons of money."

Two years and tens of billions in revenues later, Google says it has won the better part of the fight. Citing the 0.02% figure, Google appears to claim it has conquered click fraud. Now that's pretty amazing considering the awesome scale of revenues that flow through the AdWords platform.

The statement comes a few weeks after Google's chief Internet evangelist, Vint Cerf declared that up to 25% of all computers on the 'net are infected and exploited by bot-nets. In a session at the World Economic Forum in late January, Cerf suggested that, "... of the 600 million computers currently on the internet, between 100 and 150 million were already part of these botnets." (sitepronews, Jan 25, 2007)

In a prior investigation, Sitepronews learned of and wrote about botnets used to commit click fraud. One such network, made up of over 50,000 computers, was thought to make over $250K per week before it was shut down.

According to security experts, Google and other pay per click search advertising providers faced click bot activity on a daily basis. Though details are scant, a source has informed us that Google has recently adopted new methods of better detecting bot generated clicks.

We also know of pay per read and pay to click schemes working around the world. While we recognize the seriousness with which Google takes the issue of click fraud, we find it very difficult to believe Google has successfully disempowered a criminal industry known to be employing the resources of tens of thousands of people and tens of millions of computers.

If they have, we urge them to release data proving the case. In the absence of hard evidence, Google is asking the search marketing community and its advertisers to take it on its word, offering an absurdly paltry figure of $2,100,000 as the end effect of criminal click endeavors. That's a stretch of a proposition to consider.

Here's another thing to consider, completely by the numbers.

After meeting staff payroll, maintaining and expanding infrastructure, investing in R&D, buying a few cool companies, bankrolling its philanthropic foundation and paying oodles of income tax, Google made a net income of $3,077,446,000.

That net income provides a pool from which investors get to draw returns and is the first balance sheet entry looked at by financial analysts. When market research firms such as Outcast Inc suggest that PPC spending might decline by 1% this year due to advertisers' fears of click fraud, 1/30th of Google's net income is threatened. Yikes!

Has Google really cut the charges associated with Click Fraud to 0.02%? I have no idea. Neither does the next speculator. Nobody except Google is able to substantiate the number and given its historic reluctance to share any proprietary information, that is not likely to happen without intervention from the courts or from Congress.

The 0.02% claim is patently ridiculous and awfully fun to ridicule. In making it, Google is not lying but it is not telling the whole story either. Google does not provide nearly enough data to prove their case, asking advertisers to take them on their word and look at their ROI against other forms of mass-marketing for further guidance.

To be serious, Google flirts with monopolistic status in every field they become interested in. They run the world's information like nobody else's business. Google is bigger and more important than most levels of government. Everything they say has to be held to strict account.