Riding the Gravy ‘Trend’? Bandwagon Effect Vs. Conspicuous Adoption of Music in User-Generated Content

Problem definition: User-Generated Content on platforms like NetEase and TikTok serve as both means of entertainment and promotion vehicles for the music and the artists featured in them. In this paper, we study the music adoption preferences of NetEase users who actively engage in this content-making process as the function of their click-through activity while they browse the two-sided platform, and their standing in the social hierarchy that organically emerges in the platform's network structure.Methodology: We develop a theory grounded in the propensity of the users to either showcase bandwagon effect by conforming to the crowd opinion or to conspicuously deviate from the trends in their adoption strategies which is driven by their differing positions in the hierarchy. To validate the theory, we carry out our analysis on a detailed log of user activity on NetEase Coud Village registered during the month of November 2019 with a two-stage Heckman selection correction model. We first map the reward structure prevalent on the platform by analyzing the reactions that the users register for the content that they come across in the form of 'like', 'comment' and 'share'. We then analyze their music adoption strategies for the new content that they produce. Results: We find that the users stifle creativity by registering better reactions for the content that features music that conforms to the trends. Despite the reward structure that clearly favors less creative content, we further demonstrate that the socially elite users showcase a higher propensity to deviate from the trends and make the conspicuous choice of adopting music creatively in the content that they produce.Managerial implications: By applying our results to a subsample of 1,000 users to get an out-of-sample validation for user reaction and prediction for their adoption patterns, we illustrate how our framework can be utilized by both the platform and record labels in making user-engagement predictions