With the rapid development of social media, online video platform market is growing rapidly. Video analytics are becoming increasingly in demand as they help identify viewership patterns, which enable the solution provider to better understand the behavior of viewers in order to cater to their inclinations. On the basis of homogeneity theory, this paper analyzes the effect of users’ similarity network and online social interaction on consumer viewing demand. Based on data from aiqiy.com, we build the dynamic reviewer networks considering their preference similarity. The result reveals that both the density and connection intensity of reviewers' similarity network have negative demand effect, while the average node degree has positive impact. We found empirical evidence that the homogeneity of reviewers has negative effect on video views. Furthermore, we explore the comments generated from reviewers’ interaction. We find that the diversity of comments and the volume of comments have positive demand effect.