Quantifying social influence in an online music community

This paper studies two types of social influence in an online music community: observational learning influence based on aggregate consumption data, and social network influence based on music consumption by friends in social proximity. The analysis uses a variety of empirical methods, applied to highly granular user listening and "favoriting" behavior on the largest music blog aggregator site. Our analysis finds positive evidence for observational learning effects, but no evidence for social network influence. Thus, any social influence in this music context is channeled through popularity cues offered by aggregate consumption statistics, rather than contact and communication with friends in close social proximity. We discuss implications of these results for research and practice.