• by Laurie Sullivan  @lauriesullivan, May 8, 2009

Click Forensics today released an updated traffic quality management platform that allows ad networks, publishers and advertisers to identify online traffic more likely to convert from clicks into sales.

The platform aims to give customers a better way to target high-quality online traffic to improve campaign performance. Originally, the company focused on identifying bad traffic or click fraud. But the company's technology has advanced, and the data set has become large enough to identify good traffic and provide more predictive services.

The company churns out the analysis by processing billions of clicks that provide a view of traffic from more than 300 ad networks and major search engines, according to Tom Cuthbert, president of Click Forensics, Austin, Texas. "We also have advertisers that allow us to see what converts on their site and what does not," he said. "That conversation data is useful in tuning our engines in this predictive modeling system."

Some capabilities available in the platform include Click Forensics Site Score, the ability to tag traffic with individual scores that rank visitors based on their likeness to convert. This allows ad networks to filter and route traffic for the best return on investment (ROI).

Adaptive Intelligence helps the system adapt to the geographic area and traffic to give the best possible scores as information is processed. For example, an ad network that caters to retailers in Europe might have different traffic patterns from ad networks in North America.

The traffic scoring engine also taps machine-learning features that allow it to adapt in real-time to detect new sources of good and bad traffic as they emerge. Filtering decisions are made instantly, based on traffic quality thresholds set by clients.

Anomaly detection distinguishes complex click types, either made by machine or humans. This includes increasingly malicious Javascript programs that execute upon a page view or site visit. Volume and spike anomalies are more easily identified to protect from stealth attacks, such as those resulting from botnet activation.

Cuthbert said the company will continue to work on technology that predicts conversions. "The future is to find ways that leverage the information in the community, making sure advertisers get what they paid for," he said. "Advertisers want specific targets. They don't want to pay for clicks outside a certain geo-area or ones that have no chance of converting."