Read the original article in the Wall Street Journal here.
Twitter Facebook and their ilk have spawned many firms that try to extract insights from the flood of posts. One of the most ambitious is Tellagence.
The Portland, Ore., startup says it can predict how information will flow on social networks. The goal is to allow companies to figure out who is really likely to spread their messages–and avoid wasting time and money on reaching others.
Tellagence, which is giving the first public demonstration of its technology at the Demo conference on Wednesday, says it does not merely analyze what people are saying or count their followers. Nor does it try to make predictions based on past patterns of activity.
Neither approach is sufficient, the company says, since subtle and complex contextual factors affect the credibility of people who post on such networks. Moreover, many of those factors are changing all the time.
“We are not looking for the loudest or the celebrities, but the people with the right relationships,” says Matt Hixson, CEO and co-founder of Tellagence.
People’s relationships help determine who will pass messages to whom, who will listen to those messages and potentially take action on them, Hixson says. If, say, a person is not considered knowledgeable or well-connected or influential on a particular topic other people may tune out. Tellagence has developed a series of algorithms designed to help identify such roles and relationships analyze people’s behavior, providing insight to customers.
One Silicon Valley chip maker, for example, wanted to learn more about whether and how a information about a specific subject was flowing, Hixson says. By using the company’s technology, he says, the firm learned of a small group of engineers in Japan that was closely following the technology–and one specific person in the group that seemed to shape information flow about that topic.
Through such discoveries, Tellagence expects to help people identify particular people to cultivate and share information with, and avoid costly, untargeted messages.
Tellagence’s other co-founder is Nitin Mayande, who serves as its chief scientist and worked on many of the underlying principles of the technology for the past seven years as part of a PhD thesis. The resulting Web service is designed to provide customers a kind of constellation-style graph of people and links between them–which users can manipulate to see different perspectives–as well as a tactical plan for how to improve their message flow.
“What we are predicting is how far messages will travel and what path they will take,” Hixson says.
Though focusing on Twitter conversations initially, Tellagence ultimately hopes to apply its technology to other kind of business-oriented social networks such as those offered by Jive Software or Yammer, Hixson says. The company says its technology is in beta tests with “several major enterprise companies” and agencies that deal with social media.