| Abstrakt: |
With a growing number of global citizens connected via social networks, there is an unlimited potential for understanding of causal relationships that drive the spread of products, services, and information over these social networks. It is inferred that in online social networks, user characteristics and behavior tend to cluster both in space and in time. The most cited reasons are peer influence (causes her online friends to undertake a certain action) and homophily (an individual tends to befriend peers that are similar to her on several characteristics and possibly the environment they face). The peer influence can also bring about the social multiplier effect. The present research work is based on an earlier research by Aral and Walker (Ref. 1), which demonstrates how social contagion can be created by embedding viral features into product design. The present using a controlled randomized experiment in a large online social network has tried to examine whether peer influence in the general population of users exists in the context of an economic decision involving real dollars. This is studied in freemium, music-listening social network called Last.fm. The goal of freemium communities is to convert free users into premium subscribers. In order to overcome the limitations of the data that was available, the researchers have employed the gold standard of randomized controlled trials to test the hypotheses. The results are discussed on the basis of a nonparametric inference procedure and several robustness checks were employed. The results showed that the odds of a user adopting the paid subscription increase by more than 60% due to peer influence. It was also found that peer influence is weaker for users having a large number of friends. When peer influence was compared with homophily, though premium subscriptions were a rare event in the context of Last.fm, a 3% of premium users are very valuable and they contribute to more than 18% of the site’s total revenue. [ABSTRACT FROM AUTHOR] |
