Click fraud rose in Q1, both compared to a year earlier and compared to Q4 2009, according to Click Forensics Inc., which audits traffic quality for online advertisers and networks. In Q1, 17.4% of paid clicks on online ads were fraudulent, up from 15.3% in Q4 2009 and 13.8% in Q1 2009, the firm’s quarterly Click Fraud Index reports.
The fluctuation from quarter to quarter in the click fraud rate over the past three years stems from several variables including seasonality, economic climate and the technology available to perpetrate fraud.
For instance, Click Forensics reports that the proportion of click fraud attributed to botnets and malware has grown each quarter for the past three years, now accounting for nearly half of all invalid clicks. A botnet is a group of computers controlled and manipulated by one central source, often used by criminals to commit online fraud. Malware is software that is loaded onto a computer without the owner’s knowledge and with malicious intent.
The quarterly increase in click fraud in Q1, however, is somewhat surprising, the report notes. That’s because Q4 typically posts high click fraud rates because more criminals click on pricier keywords over the holidays. Click fraud rates then typically fall in Q1, the firm says.
One reason for the change could be the increase in sophisticated forms of fraud, the report says. For example, in Q1 Click Forensics saw a greater percentage of collusion fraud, which occurs when criminals try to mimic valid traffic patterns by spreading botnets or malware across a wide range of computers which then click on ads in a seemingly random pattern.
For the first time in its click fraud report history, Click Forensics measured the quality of traffic from popular social networking sites. The firm used data from third-party ad networks and directly from advertisers, not from the social media sites themselves, to determine how much of a social network’s traffic was invalid. Click Forensics detects and reports data for more than 15 different types of what it calls invalid clicks, which range from simple repeat clicks (invalid, but not fraudulent) to anomalies in volume or distribution.
The firm measured invalid clicks from Bebo, Facebook, Friendster, Hi5, LinkedIn, MySpace, Orkut, PerfSpot and Twitter. The average invalid rate for all the networks was 11.5%, significantly lower than the overall industry click fraud rate of 17.4%. Click Forensics says the reason for that lower rate is because the only motivation for generating invalid clicks on a social network’s ad would be to deplete the budget of a competitor in order to get better ad placement. A criminal can’t collect fees from such clicks, such as with paid links.
“The most common incentive for click fraud-making money by clicking on paid links to generate revenue-sharing fees-does not exist within the confines of a closed social network,” the report notes.
Additionally, because social networks require users to log in, perpetrators of fraud can be identified and social networks can shut those accounts down quickly.
Below are the Q1 click fraud rates since 2006, when Click Forensics began publishing industry data:
- 2006: 14.2%
- 2007: 14.8%
- 2008: 16.3%
- 2009: 13.8%
- 2010: 17.4%