To Partner or Not to Partner? The Partnership Between Platforms and Data Brokers in Two-Sided Markets

• 成果类型: Article
• 署名作者: Zhang, Xin; Cheng, Lihong; Yu, Yugang; Tan, Yong
• 署名单位: Chinese Academy of Sciences; University of Science & Technology of China, CAS; University of Washington; University of Washington Seattle
• 刊物名称: INFORMATION SYSTEMS RESEARCH
• ISSN/ISSBN: 1047-7047
• DOI: 10.1287/isre.2022.0470
• 发表日期: 2025
• 关键词: user privacy concerns; COMPETITION; ECONOMICS; INFORMATION

摘要

Data have become an important competitive asset, especially for online advertising platforms, leading to the emergence of a new data broker industry. Platforms can partner with data brokers to acquire external data to enhance their targeting capabilities, but this practice has stoked growing consumer privacy concerns. This study develops a gametheoretic model to examine the economic mechanism underlying the partnership between competing platforms and a data broker in a two-sided market. Interestingly, our analysis shows that increasing consumer privacy concerns caused by the data broker may encourage platforms to partner with the data broker rather than discourage them. The driving force is that, because of negative cross-side network effects from advertisers to consumers, increased privacy concerns can serve as a strategic lever to soften price competition on the advertiser side. However, when both platforms partner with the data broker, a prisoner's dilemma may arise in equilibrium even when the data broker can greatly improve the targeting capabilities. We find that the platform-data broker partnership hurts consumer surplus when platforms adopt a pure ad-sponsored model without charging consumers, but it may benefit consumer surplus when platforms adopt a mixed model with ad-sponsored and subscription-based revenue. Moreover, our analysis indicates that platforms may not have incentives to adopt high levels of privacy protection and may instead adopt asymmetric privacy protection strategies, with one platform adopting high privacy protection and the other opting for low protection. A series of model extensions confirm the robustness of our key findings. We conclude with a discussion of managerial and policy implications.