Everyone Can Be a Star: Understanding the Role of Live Video Streaming in Online Retail

Abstract

With the rising popularity of live video streaming in recent years, it has begun to attract interest from both e-commerce practitioners and the research community. However, there is scarce academic research in understanding whether and how live video streaming affects the landscape of e-commerce. In this paper, we seek to understand the effects of the adoption of live video streaming strategy by online retailers and their online product sales. Leveraging multiple archival data sources, we combine econometrics with deep learning to test the hypotheses. The results indicate that adopting live video streaming strategy significantly boosts online product sales. Furthermore, we propose a novel deep learning framework to address the challenge of identifying product-oriented live video streaming, and find that the adoption of a product-oriented live video streaming strategy increases online product sales, whereas there is no significant impact for adopting non-product-related live video streaming strategy. In the meanwhile, the impact of product-oriented live videos is attenuated as the buyers’ uncertainty decreases. These results provide empirical support that reduction in product uncertainty via information provision is the main mechanism through which live video streaming increases online sales.

Keywords: live video streaming, online retail, business value of IT, difference-in-differences, look-ahead matching, deep learning, video classification, multimodal learning