Impact of data policy on platform economy

Abstract

The proliferation of data-driven personalization in digital platforms has become increasingly critical in today’s platform economy. This focus on leveraging user data plays a pivotal role in enhancing user engagement and optimizing platform service quality, while navigating the complex landscape of user privacy concerns. As users’ awareness of data privacy increases, regulatory policies represented by GDPR require individuals to have greater control over their personal data, which affects the platform’s data collection strategies. However, implementing effective data policies is not straightforward. Platforms face distinct challenges and trade-offs when balancing personalization benefits with privacy protections. In practice, we observe three primary data collection scenarios: (1) Data-Free Collection (N), where platforms rely on generalized recommendations without user data; (2) Mandatory Collection (M), where platforms collect comprehensive user data for personalized recommendations; and (3) Voluntary Collection (V), where users opt-in to data sharing, balancing personalization with privacy. In this thesis, we aim to investigate the impacts of different data collection scenarios on a monopolistic platform’s recommendation strategies, pricing decisions, and user demand. We analyze the platform’s decisions and performance under each data collection scenario and evaluate their implications for user surplus and platform profitability, focusing on two main chapters: the impact of data policies on user polarization and platform service level. Our analysis yields several key insights. (1) First, the effectiveness of recommendation strategies and user polarization tolerance significantly influence the platform's optimal data collection choice. In scenarios with high privacy sensitivity, Scenario N may outperform by leveraging generalized recommendations, while Scenario M excels in contexts where personalization drives engagement. Scenario V often balances these trade-offs, maximizing consumer surplus by aligning personalization with user consent. (2) Second, Scenario V frequently achieves a win-win outcome by enhancing service quality while respecting privacy preferences. (3) Third, higher recommendation accuracy under Scenario M may not always benefit users if privacy concerns increase searching costs, whereas Scenario N can mitigate this in privacy-sensitive markets by reducing the searching cost. Our results provide actionable managerial insights for platform operators, emphasizing the need to carefully select data collection strategies that align with user preferences and market conditions to optimize engagement and profitability.