Dynamic Competition Identification Through Consumers’ Clickstream Data

Abstract

Brands that use online marketplaces face challenges on identifying the market structure and analyzing their competitiveness. We address that lacuna by modeling online consumers’ behavior using clickstream data and considering the interdependence of brands using network analysis. We draw on a dataset of 6,549,484 records over a period of 10 weeks from one of the biggest online marketplaces in China and employ spatial auto-regressive models and network structural properties of brands to predict sales. Our findings indicate that intra-brand competition is more intense than inter-brand one and is the main reason for the fluctuations of sales. Concurrently, we demonstrate the redistribution of market shares of related products after the firm adjusts the length of the production line, so as to provide a reference for how to adjust the length intra-brand. By exploring the relationship between the structural position in the network and brand sales, we show that the span of structural holes of a brand negatively influences sales, while betweenness and degree centrality has a positive impact on sales. Our study contributes to the better understanding of brand competition on online marketplaces and presents both theoretical and practical implications. We discuss the significance of our findings for brand competition on online marketplaces and platforms, while we draw an agenda for future research on the topic.