Na Liu
Supervised by Prof. Chris Forman (Cornell University)
R&R, Information Systems Research
This study examines how commission fee pricing reshapes multi-sided user-generated content (UGC) platforms by analyzing a natural experiment where a major platform eliminated fees for goods-selling ad links in sponsored contents. Using various market-level difference-in-differences models, we reveal three key findings: (1) Advertisers reallocated 13% of placements toward fee-exempt links, demonstrating notable price sensitivity; (2) Platform revenue declined as substitutions outweighed demand expansion; and (3) The fee exemption triggered a redistribution on the creators' side, with contents for treated advertisers gaining three additional interactions per 100 views, occurring primarily from top creators. Our work advances information systems research by providing the first causal evidence on how commission fee changes reconfigure multi-sided markets—spanning advertisers, creators, and platform revenues. The findings highlight critical trade-offs: while fee structures steer advertiser behavior, they may inadvertently cannibalize revenue and amplify redistribution among participants in this multi-sided ecosystem. These insights inform platform governance by demonstrating the need to balance participation with profitability and model cross-sided effects before implementing pricing changes.
This study examines how commission fee pricing reshapes multi-sided user-generated content (UGC) platforms by analyzing a natural experiment where a major platform eliminated fees for goods-selling ad links in sponsored contents. Using various market-level difference-in-differences models, we reveal three key findings: (1) Advertisers reallocated 13% of placements toward fee-exempt links, demonstrating notable price sensitivity; (2) Platform revenue declined as substitutions outweighed demand expansion; and (3) The fee exemption triggered a redistribution on the creators' side, with contents for treated advertisers gaining three additional interactions per 100 views, occurring primarily from top creators. Our work advances information systems research by providing the first causal evidence on how commission fee changes reconfigure multi-sided markets—spanning advertisers, creators, and platform revenues. The findings highlight critical trade-offs: while fee structures steer advertiser behavior, they may inadvertently cannibalize revenue and amplify redistribution among participants in this multi-sided ecosystem. These insights inform platform governance by demonstrating the need to balance participation with profitability and model cross-sided effects before implementing pricing changes.
Intermediaries in the UGC Digital Economy: MCN Roles and Revenue Sharing
Supervised by Prof. Chris Forman (Cornell) and Prof. Michael Zhang (CUHK)
Multi-Channel Networks (MCNs) have become central intermediaries in user-generated content (UGC) platforms, yet their strategic behavior under platform incentive programs remains poorly understood. Using a natural experiment on a major video platform that offered bonus traffic to MCNs for signing new Dance creators, we combine difference-in-differences (DiD) and triple-difference (DIDID) designs with an agent-based simulation. We find that although the policy aimed to support new creators, MCNs reallocated the bulk of bonus traffic to large incumbent creators. Category-level intent-to-treat (ITT) effects are null, indicating redistribution rather than expansion. The simulation shows that MCNs rationally favor large creators to avoid high defection risk, even at the cost of short-run efficiency from negative marginal GMV gains. Our results contribute to platform governance, intermediation theory, and the creator economy literature by demonstrating that delegated incentives can produce unintended consequences when intermediaries pursue profit-maximizing retention strategies.
Consumers’ Preferences and Aversion in AIGC
with Prof. Catherine Tucker (MIT)
AI-Generated Content (AIGC) is rapidly transforming digital economies, yet its impact on consumer behavior and trust remains poorly understood. This study investigates how transparency around AI use influences user engagement and economic decisions, with a focus on distinguishing intrinsic aversion (a “taste” for human-made content) from instrumental aversion (fear of economic loss). We design a large-scale field experiment implemented via a retail platform and ad campaigns, testing three disclosure conditions using click-through and conversion rates as behavioral metrics, we quantify the two forms of AI aversion and examine their underlying mechanisms. We supplement experimental results with heterogeneity analysis across demographic and industrial segments, mechanism surveys identifying drivers of trust erosion, and a computer vision model trained to detect AIGC content. Furthermore, we explore the use of GPT-4 as simulated respondents to assess its viability for experimental information system (IS) research. Our findings provide nuanced insights into how AI shapes consumer preferences and platform outcomes, offering strategic guidance for firms and policymakers aiming to implement ethical and effective AIGC deployment
Firm Heterogeneity and AI Industry Development
with Dr. Ting Ma (The Institute of Scientific and Technical Information)
Preparing Submission to Research Policy
Preparing Submission to Research Policy
Market Structure and Location-based Food Deserts in Smart Cities
with Prof. Nathan Yang (Cornell)