Predictive Networks Invention Uses Biometrics

Predictive Networks announced the company has filed an application with the United States Patent Office to patent its method for uniquely differentiating individuals in households in which devices such as computers and interactive televisions are shared by multiple users. The invention combines network-based clickstream data with "biometrics," based on each user's keystroke and mouse or remote control usage patterns, and uses no personally identifiable information - only randomly-generated, anonymous IDs. The approach represents the first use of such biometric data for auto-personalization of content and advertising, solving a major challenge facing marketers on the Internet, as well as wireless networks and interactive television (iTV), where devices are intrinsically shared.

Predictive Networks is the creator of artificial intelligence-based Digital Silhouettes -- detailed, but anonymous portraits of users based on their behavior in cyberspace. Individual users manifest different surfing styles, preferences and affinities, but also have different typing styles, as well as mouse usage patterns. By creating a "fusion algorithm," incorporating data points from clickstreams across the network, along with biometric information from keyboard and mouse ore remote activity, the company can pinpoint individuals and help advertisers, commerce and content providers target their offerings on-the-fly. Privacy is protected by using only random ID numbers -- not personal information -- by discarding all clickstream data after analysis, and by advancing beyond "opt-in" approaches to provide users with complete control over their identities in cyberspace.

"Biometrics give us the answer to a difficult question - which user has the controls?" said Devin Hosea, Predictive Networks president and founder. "Initially, we want to use this technology to help our partners better personalize their offerings by uniquely differentiating users. In the next phase of its implementation, we'll be looking at building correlations between biometric signatures and our Digital Silhouette affinities, making it much faster to auto-personalize when a new user comes to a device, such as a kiosk, or borrows a computer in someone's home or office. This technology gives us the unique ability to help our partners personalize their offerings across numerous, casually-accessible devices."

About Predictive Networks

Predictive Networks provides network infrastructure for targeting and personalization across the Internet, interactive television and wireless networks. By combining its patent-pending, highly characterized cyberidentities with privacy technology that empowers consumers, the company enables content providers, iTV platform providers, portals, network and service providers, and digital marketers to deliver more personalized experiences and more relevant offerings to their users, while protecting individual privacy. Privately held and founded in 1999, Predictive Networks is backed by strategic corporate investors, including Battery Ventures, Advent, NTT, Net2Phone, NetRatings, and Unterberg Towbin. For more information, please call 617-575-4700 or visit our Web site at www.predictivenetworks.com.

Contact Predictive Networks, Betsy Kosheff, 617-575-4614, bkosheff@predictivenetworks.com.