Sunsports Lead to More Successful Searching

E-commerce may be a war in progress, but who would have guessed that the next weapon in the online marketer's armory would come from real-world battlefields?

In the early 1990s, George Roumeliotis was a senior research scientist in the astrophysics group at Stanford University, where he was studying solar flares. These violent phenomena on the sun's surface can have detrimental effects on earthly telecommunications--a subject of great interest to the U.S. Air Force. Roumeliotis was specifically trying to find ways to predict the occurrence of flares before they happened--giving the military a heads up that battlefield communications could be about to go on the blink.

Scientists studying the flares often had sparse data to interpret--perhaps only a few photons in a given frequency range, Roumeliotis says. Consequently, the researchers developed pattern recognition algorithms capable of producing reliable results but with minimal information.

Then, in the late 1990s as the Internet boomed, Roumeliotis made an intuitive leap: If such algorithms could predict potential explosions on the surface of the sun based on limited data, why couldn't they be used to predict the behavior of online shoppers based on observations of their actions?

To test the theory, Roumeliotis and his partners formed Dynaptics Corp. (www.dynaptics.com). The Cupertino, Calif.-based company makes products that rely, at least in part, on the earlier solar-flare research. According to Roumeliotis, Dynaptics' personalization products offer several advantages over traditional search and hard-coded online tools. These products, called E-ssistants, don't require much information to get started. They don't make the mistargeted offers that hard-coded "hit this page; get this suggestion" tools often provide.

And E-ssistants don't require historical data--they follow only a single user session rather than setting cookies and tracking long-term behavior. That could be important if online privacy issues begin to make cookies and other monitoring tools more difficult to use without explicit permission.

Dynaptics recently unveiled the Personal Shopping E-ssistant and the Personal Information E-ssistant, which work in different ways to achieve the same end--increased purchases by online shoppers. The Shopping E-ssistant tracks Web visitors' movements and targets additional information at them--perhaps a suggestion for a related product or a special-package discount (offering a helmet to go along with a bike, for example)--depending on the choices they make. The Information E-ssistant helps buyers dig up products they might otherwise miss at auction and exchange sites.

• Bookmark this page
• Share this article
  •  
  • facebook
  • slashdot
  • digg
  • Reddit
  • stumbleupon
  • linkedin
  • twitter
• Got more on this story? Email Computerworld
• Follow Computerworld on twitter

• Share
  Share this article

  •  google+
  • facebookfacebook
  • slashdotslashdot
  • diggdigg
  • RedditReddit
  • stumbleuponstumbleupon
  • linkedinlinkedin
  • twittertwitter
• print
• email
• Bookmark