OIM Faculty Research: Online Rating Systems and Fake Social Media Engagement

By Ying Liu, assistant professor

As digital platforms increasingly influence consumer decisions, understanding how users interpret online information is vital for businesses and policymakers. My recent research explores two related topics: how multidimensional rating systems affect consumer satisfaction, and how fake social media engagement (astroturfing) shapes market outcomes and consumer behavior.

The first study, published in Management Science, examines the value of multidimensional (MD) rating systems compared to single-dimensional (SD) systems. While many review platforms still rely on a single overall score (e.g., Yelp stars), MD systems allow users to rate different dimensions, like food, service, and ambience separately. We analyzed a natural experiment in which TripAdvisor introduced MD ratings, using Yelp as a control group in a difference-in-differences design. Our findings show that after the introduction of MD ratings, TripAdvisor’s average ratings increased by 0.4 stars and became more consistent over time. Consumers using MD systems made matched decisions and reported higher confidence in their choices. Follow-up randomized experiments supported this mechanism: MD systems helped reduce product uncertainty and improved satisfaction. These results suggest that better rating designs can lead to more effective decision-making for users.

The Hidden Costs of Social Media Manipulation

In contrast, my second research project, published in the Journal of the Association for Information Systems, examines a more concerning phenomenon: social media astroturfing, the use of fake posts or automated accounts designed to artificially generate buzz and perceived popularity. Using panel vector autoregression (PVAR) and a dynamic matching method, we analyzed the impact of fake posts and accounts on movie performance. We found that social media astroturfing can temporarily boost box office revenues, particularly if used before a film’s release. However, these effects quickly fade. More importantly, movies that used astroturfing saw declines in user ratings over time, indicating that consumers experienced a gap between the promotional buzz and their actual viewing experience. This not only harms brand reputation but also undermines trust in social media platforms. Our analysis further shows that astroturfing distorts the feedback ecosystem, drowning out authentic voices and misleading both consumers and competitors.

Taken together, these two studies offer contrasting perspectives on digital influence. Thoughtful system design, such as MD ratings, can enhance transparency and user satisfaction. In contrast, deceptive tactics like astroturfing may offer short-term gains but carry long-term reputational risks. For alumni working in tech, marketing, or analytics, these insights highlight the importance of building trustworthy digital systems and being cautious of short-term strategies that could damage user confidence.

As digital ecosystems evolve, I hope this research contributes to ongoing conversations about how to balance innovation, integrity, and consumer welfare in the digital environment.

About Ying Liu, PhD

Ying Liu is an assistant professor in Operations & Information Management at the Isenberg School of Management at UMass Amherst. She holds a PhD in computer information systems from the W.P. Carey School of Business at Arizona State University as well as a bachelor’s degree in electronic information engineering from the School of Telecommunications at Beijing University of Posts and Telecommunications.

Liu’s research focuses on information economics, digital marketing, user-generated content, social media and social networks, and crowd marketing.

She teaches Machine Learning in Business and Machine Learning for Analytics.