Predicting Membership-Sites' Success

Guest_Jim_* - February 4, 2014 01:26PM in Science & Technology

Why do some websites succeed and others fail? This question has been plaguing many people, especially those who build the websites and invest in them. Researchers at Carnegie Mellon University have recently developed a new model for predicting the success of membership-based sites, such as Facebook.

Modelling membership sites is a different undertaking than modelling other systems as instead of possessing a limited amount of information, there is an overabundance and attention is what is limited. This creates a "marketplace of attention," as the large Nobel laureate Herbert Simon called it. From Simon's writings, the Carnegie Mellon researchers were inspired to consider the many dynamics involved in a site's success, such as active users becoming inactive, and users getting friends involved.

Data on daily active users for the past six years were obtained for twenty-two sites from the Web analytics company, Alexa. Facebook was among these sites, and according to the data and the new model, it should be sustainable for the foreseeable future, but of course the model cannot predict an upstart site.

Source: Carnegie Mellon University