Information Needs in the Everyday Usage of Video Social Recommender Systems

Abstract

This thesis investigates the information needs of users in the everyday usage of Video Social Recommender Systems (VSRS), with a focus on understanding the motivations and contexts of such usage. Utilising a diary-based qualitative approach, this study captures real-time data from users to minimise recall bias inherent in traditional retrospective methods. The research employs Taylor’s levels of information needs to analyse the data and explores the concept of serendipity as a design principle for VSRS. Key findings reveal that user motivations are predominantly emotive, with VSRS serving as a tool for relaxation, distraction, and entertainment. Moreover, the study uncovers habitual usage patterns, particularly around sleep-related contexts, highlighting the integration of VSRS into users’ daily routines. The absence of Q4-compromised needs as initial interactions emphasises the serendipitous nature of VSRS, where users rely on the algorithmic recommendations rather than explicit searches. The results suggest a potential commodification pitfall, where users have implicit expectations of the system to fulfill diverse needs without direct input. This underscores the importance of designing VSRS that effectively balance user engagement with unexpected yet valuable content encounters. Overall, this research contributes an initial information science perspective on the complex dynamics of VSRS usage. Future research avenues include exploring the serendipitous expectations and flow experiences within VSRS interactions.