ValueClick Tests Predictive Analysts In Behavioral Targeting

by Laurie Sullivan, Mar 19, 2008, 7:00 AM

• Comment
• Recommend (4)
• Tweet
• 

ValueClick Media said Tuesday at the OMMA Hollywood conference that it has built an applications platform to offer advertisers behavioral targeting based on predictive analytics. The company will officially launch the offering this summer, but has been working with five unidentified mobile service providers in a private test since Feb. 1.

The mobile service providers have been targeting consumers identified through cookies that live on PCs. The data fed from the cookies into algorithms suggest that these visitors will more likely respond to mobile offers.

This week, ValueClick opened the beta to more advertisers in the company's newsletter, saying the new behavioral application promises to deliver improved performance to marketers. By focusing on targeting visitors rather than pages, ValueClick says it can serve more relevant ads that appeal to each Web site visitor based on past behavior to provide higher click-through and conversion rates.

ValueClick's platform relies on three algorithms: behavioral targeting, recommendations and ad server real-time selections, according to Joshua Koran, VP of targeting and optimization at ValueClick, which supports media inventory on more than 13,000 sites. "Cookies stored on visitors' computers collect data and build search profiles containing one or more facts," he said.

The cookies collect data on Web site visits, ad interactions, searching, product comparisons, product purchases, and third-party data fed into algorithms that predict behavioral trends, storing the information in ValueClick servers. Koran says the predictive analytics platform will cluster the data collected from searches and clicks in thousands of categories, such as mobile, travel and finance.

The predictive algorithms in the behavioral targeting application will store the profile as long as the cookie comes back to the server with more data. It will process hundreds of millions of profiles tied to anonymous cookies on billions of events daily. The algorithm built into the application will determine when the information is no longer useful.

Still, relying solely on cookies to collect consumer data isn't foolproof. Between 10% and 30% of people delete cookies on their PC. "It's an illustration as marketers we're not doing a good job in showing people it's to their benefit to share information," Koran says. "Consumers don't want to share information to protect their privacy."

Privacy remains an issue because the industry lacks clear description on how and why companies collect information, Koran says. "That's why they think Big Brother is watching them, and in some cases Big Brother is watching them, but it's not us," he said. "We plan to put policies in place that describe in plain English our intent."

• Comment
• Recommend (4)
• Tweet
• 

Be the first to comment on "ValueClick Tests Predictive Analysts In Behavioral Targeting "

Leave a Comment