No Movie to Watch: A Design Strategy for Enhancing Content Diversity through Social Recommendation in the Subscription-Video-On-Demand Service

Increasing diversity is becoming crucial in recommender systems to address the “filter bubble” issue caused by accuracy-based algorithms. Diversity-oriented algorithms have been developed to solve this problem. However, this diversification has made it difficult for users to discover what they really want from the variety of information provided by the algorithm. Users spend their time wandering around the recommended content space but fail to find content they want to watch. Therefore, they rely on external services to gather information that does not appear on the recommended list. This could lead to a reduction in the services’ ability to compete with other subscription video on-demand (SVOD) services. To address this problem, this study proposes a human-centered approach to diversification through social recommendations. We conducted an experiment to understand how perceived diversity affects user perceptions and attitudes. Specifically, by incorporating social recommendations into the SVOD service, this experiment was changed to examine the following conditions: (1) influencers vs. online friends, and (2) human recommendation lists vs. algorithmic recommendation lists. The findings indicated that perceived diversity influences the manner in which the users perceive information quality and playfulness, both of which have a positive effect on their intention to use. Additionally, the participants’ perceptions of information quality were greater in the scenario with the human recommendation than in that with the algorithmic recommendation. This study contributes to the development of a theoretical framework based on perceived diversity through social recommendations and the design of an SVOD interface with social recommendations to provide better user experiences.