UROP Proceedings 2022-23

School of Business and Management Department of Information Systems, Business Statistics and Operations Management 179 Department of Information Systems, Business Statistics and Operations Management Identifying Entrepreneurs' Storytelling Strategies Using Twitter Data Supervisor: KWON, Ohchan / ISOM Student: JIANG, Yuxuan / ECOF Course: UROP2100, Fall Twitter, as a mainstream social media, is a trendsetter for various activities or events. At the same time, its vast database supports many analyses about reaction of users on individual level. Based on the strengths of Twitter data, our study explores two different aspects: the celebrity effect of NBA players on knowledge dissemination of science, and the acceleration driving force of startup bootcamp on the success of first-stage entrepreneurs. This report serves as a progress record for exploration in these two questions and their respective research progress, including basic introduction, literature review, identification strategy, data description, and discussions for next stage. A Study on Amazon Aggregators Supervisor: KWON, Ohchan / ISOM Student: DENG, Hanwen / WBB Course: UROP3100, Fall There is a growing concern that some products on online marketplaces are counterfeits. Combating counterfeits is not easy because they are deceptive and exist in large quantities. Recently, online marketplaces such as Amazon have adopted artificial intelligence (AI) to detect and remove counterfeits. In this paper, we study the effects of Project Zero, Amazon’s artificial intelligence to combat counterfeits, on its online marketplace. Exploiting that Project Zero launched first in the United States and then in other countries, we use a difference-in-differences design comparing the same products between Amazon US and Canada marketplaces. We find that the AI technology reduced seller competition and improved customer satisfaction. The effects are more pronounced in products with presumably more counterfeits and appear to be driven by the removal of low-quality sellers. Although some theory suggests that reduced seller competition might lead to higher price levels, we observed an increase only in the products’ lowest prices but not in the buy-box prices, that is, prices from sellers that Amazon recommends algorithmically for each product. It implies that low-quality counterfeits and genuine products target different customer segments and explains why reduced competition does not necessarily damage customer satisfaction. Our results show that counterfeits lower customer satisfaction in online marketplaces and that AI can improve overall platform quality. Our research shows that the presence of low quality sellers selling counterefeits reduce customer satisfaction and that platforms can use AI to efficiently remove low quality sellers.