Managing Online Reviews for Business Impact
(A2026-133914)

AUTHORS

Leif Brandes, University of Lucerne; Tobias Schlager; Sanjana Rosario, Columbia University; Jasmine Gould, University of Luzern; Marton Varga, Bocconi University

ABSTRACT

Online reviews are an important source of information for both consumers and businesses alike. For instance, 93% of consumers say that reviews influence their purchasing decision (Podium, 2022). The overall quality of reviews is central to success: the ‘better’ the available review content, the more efficient the resulting allocation and matching of products and services to potential target consumers. But what does ‘better’ review content actually look like, and how does it impact consumer behavior and market outcomes? How can review platforms and individual businesses manage review content to impact consumer search and market outcomes? Each of the four papers in this special session elucidates one central aspect pertaining to these two questions. The goal of this session is to show how (not) to manage reviews for business impact. Specifically, we integrate papers that (1) show how to correct existing biases, (2) demonstrate the implications of review quality and business actions on consumers during product search, and (3) how platforms can make review content more impactful by using it as basis for business awards. The four papers span consumer psychology, text analysis, econometrics and digital platform design, and should appeal to a broad audience. Because of this conceptual and methodological breadth, no competitive track adequately accommodates the session. Distributing the papers across sessions would undermine the cross-paper insights and the integrated view of how to manage review content for business impact that the session is designed to generate. The session proceeds from studies that measure review quality directly to studies that focus on the impact of review quality on consumer search and business strategies. The first two papers measure the quality of various review elements directly. First, Tobias Schlager studies the use of emojis as a form of paralanguage elements and shows that imagery values from emojis increase helpfulness and purchase intention of the reviewed product. Second, Sanjana Rosario demonstrates how to reduce the extremity of online rating distributions. The paper builds on collected training data that does not suffer from self-selection bias and develops a probabilistic model of reviewing which accounts for differential propensities of users to review based on their experience. This approach allows to create a corrected distribution of reviews that does not suffer from self-selection for out of sample products. The third and fourth papers study review quality as it impacts the behavior of consumers and businesses. Jasmine Gould leverages a unique feature of a large European online travel platform to assess what consumers read. Using a large dataset consisting of clickstream data, review content, and booking information, she reports that verified reviews containing more negative information, more photos, and/or a hotel management response are more likely to be read. Finally, Marton Varga examines how review-based platform recognition affects seller behavior. Drawing on data from 13 product categories and more than 4.3 million observations from a multi-seller platform where firms offer identical products and compete side-by-side, the authors investigate the impact of an annual “Firm of the Year” award— a recognition largely based on verified customer reviews and determined by a proprietary, black-box algorithm. The authors show that recent award recipients raise prices, expand assortments, and reduce off-platform communication.