Abstract
Problem definition: Digital platforms now guide most consumption decisions through personalized assortments and ranked recommendations. At the same time, these platforms have strong self-preferencing incentives to push high-margin or strategic content (e.g., originals, sponsored items, private labels). When such incentives are salient, users may no longer treat rankings as neutral advice, but instead question their credibility and adjust their behavior. We study how this tension between platform incentives and user trust shapes engagement, assortment decisions, and long-run platform payoff.
Methodology/results: We develop a game-theoretic model that captures the strategic interactions between a platform and rational users who adjust their beliefs and decide whether to follow, discount, or ignore the recommendations. We introduce platform’s self-preferencing weight toward the high-margin category. Our analysis reveals three distinct equilibrium regimes: compliance, pandering, and zero-compliance. We find that persistent self-preferencing is not always beneficial; excessive bias can lead users to disregard recommendations entirely. Reducing assortment variety can enhance both credibility and profitability, challenging the conventional belief that “more choice is better.” Furthermore, we propose strategies such as donation-match schemes to restore user trust while maintaining profitability.
Managerial implications: Our analysis delivers concise design rules for incentive-driven recommendation systems. When users recognize platform-incentivized options, content curation and ranking should be treated as tools for credibility engineering: sometimes limiting platform-favored items, sometimes softening self-preferencing, and sometimes using donation-match type, value sharing schemes that visibly align the platform’s interests with those of its users and restore trust in the recommendation signal.
Author: Xing Hu