The user's communication patterns on a mobile social network site

Given that users are simultaneously connected in multiple communication channels in a social networking service site (e.g., chat, message, and group message), we explore user's collective networking behavior. We collected the data from a mobile social networking site with 4.8 million registered users. The empirical estimation shows interesting results: (1) there are cross-effects across the communication channels: substitute effects for "chat and message" and complementary effects for "message and group message" and "chat and group message" (2) there is significant local network effect but global network effect is not observed, (3) users utilize communication channels for different purposes according to their networking activity level (conveying simple information vs. building sophisticated inter-relationship), and (4) we identify the distinct evolutionary trajectories of an individual user's networking behavior by channel: negative slopes for chat and message vs. upward trend for a group message. Our experimental study shows that we can better predict the word of mouth (WOM) effects by understanding users' collective networking behavior across diverse channels.