An analysis of popularity information effects: Field experiments in an online marketplace

• Analyzing empirically the effects of the popularity information on sales.
• Presenting the results from controlled field experiments in an online store on Tmall.
• Identifying causality using difference-in-differences and propensity score matching.
• Presenting the hit list effects varying over subgroups of price and category.

Popularity information is identified as one of the important factors for businesses in the online marketplace and is normally expected to strengthen sales performance. At the same time, the effects of popularity information are found to vary across product types. In this study, we attempt to identify the effects of popularity information on product sales through an analysis of the subgroups of category and price. Two sequential field experiments in an online apparel store on Alibaba’s Tmall, the most dominant online brand marketplace in China, are conducted to capture the causal effects of the popularity information on sales. After observing the possible existence of popularity information effects through a pilot test of 17 products, we conduct the main experiment with 290 products, recording the daily sales for each by posting selected products on the hit list. The difference-in-differences method and propensity score matching are used to analyze the effects. The results show that once the products are displayed on the hit list, product sales increase by an average of 1.3 units per day. One subgroup of the niche product category is found to be influenced more significantly by hit list information than are other subgroups in the broad appeal category. Furthermore, after the hit list information is presented, more units of mid-price products are likely to be sold than units of products with high and low prices.