Cheolho Song

I am a Ph.D. candidate in Quantitative Marketing at Olin Business School, Washington University in St. Louis. I am advised by Tat Chan and expect to complete my Ph.D. in December 2026.

I am on the 2026–2027 academic job market.

My research develops and applies econometric methods to study how marketing actions and consumer information affect demand on retail platforms and in other markets with multiple sellers. My current empirical domains are retail promotions and electronic word of mouth (eWOM), where I study promotion effects, customer acquisition and reactivation, and spillovers across sellers. Methodologically, I use causal inference, machine learning, and structural modeling, choosing the specific tool to match the target estimand and identification strategy in each project.

Quantitative Marketing Retail Platforms Retail Promotions eWOM Causal Inference Causal Machine Learning Structural Models Machine Learning

Education

  • Washington University in St. Louis
    Ph.D. Candidate in Quantitative Marketing, expected December 2026
  • Seoul National University
    M.S. in Business Administration, 2020
  • Seoul National University
    B.A. in Economics, 2016

Research

Whether, What, and How Much to Buy: A Semiparametric Framework for Estimating Promotional Effects on Retail Platforms

Job Market Paper · with Jingcun Cao, Tat Chan, and Dennis Zhang

Abstract. This paper develops a semiparametric framework for estimating the effects of promotions on a retail platform on consumers’ decisions of whether to visit the platform, what product categories to purchase, and how much to buy. Our framework flexibly models these multi-stage correlated decisions, allowing the use of high-dimensional individual shopping history variables to capture heterogeneous shopping patterns across consumers. We use Neyman-orthogonal scores with nested cross-fitting to debias the estimated effects and establish √n consistency and asymptotic normality of the estimators. We apply the framework to mall-wide promotions in a luxury shopping mall in China. The promotions broadly increase consumers’ mall visits, category purchases, and total spending across all store categories, with luxury stores generating a large portion of the gain, but negative effects exist in the pre- and post-window, indicating a partial intertemporal substitution. We also find that, during promotions, consumers significantly increase spending in store categories in which they spent the most before the promotion; however, intertemporal substitution effects are also the largest in these categories. By contrast, spending also increases in other store categories with little decline in the pre- and post-window, suggesting spending expansion beyond customers’ usual categories.

Celebrating the Shopping Festival: The Effects on Retail Sales and Its Drivers in a Shopping Mall

Under revision at the Journal of Marketing Research (risky revision) · with Jingcun Cao and Tat Chan

Abstract. This paper compares coordinated mall-wide shopping festivals with individual in-store promotions using nine years of transaction data from a large shopping mall in China. Both types of promotions increase sales and purchasing traffic, but shopping festivals generate substantially larger effects and draw responses broadly across customer segments, with the largest relative gains among customers who are new to the mall or have not shopped there recently. In-store promotions, by contrast, generate stronger responses among customers with prior shopping history at the promoted store. A theoretical model explains both patterns through differential promotion salience: store-specific promotions make the focal store especially salient to store-familiar customers, whereas mall-wide festivals make the mall salient to a broader set of consumers. The findings show how a central platform can coordinate promotions across sellers to acquire new customers, reactivate dormant customers, and generate demand beyond what individual stores can achieve.

Spillover Effects of Online Reviews: Evidence from the Hotel Industry

Working Paper · with Hülya Karaman (equal authorship), Tat Chan, and Song Yao

Abstract. Do online consumer reviews have spillover effects on the sales of competitors? We study this question using a natural experiment where a major hotel group introduced its own review system to all the hotels within its portfolio. Using a modified Synthetic Control Method, we analyze the monthly financial performance data of reviewed and competing hotels before and after the review system was introduced. We find that the review system has significant economic impacts on both reviewed and competing hotels, and that the effects are highly heterogeneous. Surprisingly, the correlation between the economic impact on reviewed and competing hotels is significantly positive: If the occupancy rate of a reviewed hotel increases by 1%, its competitors’ occupancy rate increases by 0.39%. The positive correlation can be explained by the information spillover. Under this mechanism, consumers update their beliefs about competing hotels based on the reviews for reviewed hotels. We explore alternative explanations, but none can fully explain the positive correlation. Our results suggest that, besides monitoring their own reviews, managers should also monitor competing products’ reviews because of information spillover from the competitors’ reviews.

Conference Presentations

  • The Optimal Design of Platform-Wide Conditional Promotions: An Empirical Study
    Earlier version of the job market paper · Annual AI in Management Conference, USC, 2025; ISMS Marketing Science Conference, Sydney, 2024; China India Insights Conference, The University of Hong Kong, 2024
  • Celebrating the Shopping Festival: The Effects on Retail Sales and Its Drivers in a Shopping Mall
    China India Insights Conference, Stanford University, 2023
  • Spillover Effects of Online Reviews: Evidence from the Hotel Industry
    ISMS Marketing Science Conference, Miami, 2023; Graduate Research Symposium, Washington University in St. Louis, 2023
  • Comparison of the Characteristics of Organic Consumers in Korea by Sales Channels
    Association for Consumer Research Conference, Orlando, 2015
  • Social Media Opinion Mining Analysis by a BI Application and New Product Sales Performance
    Post-ICIS LG CNS/KrAIS Workshop, Milano, 2013

Teaching

Teaching Assistant — Washington University in St. Louis

  • Empirical Methods in Structural Modeling (Ph.D.), Spring 2025
  • Empirical Methods in Business (Ph.D.), Fall 2022
  • Text Mining, Spring 2024 and Spring 2025
  • Customer Analytics, Fall 2021, Spring 2023, and Fall 2024
  • Analytics-Driven Brand Management, Spring 2022
  • Data Analysis for Brand Management, Spring 2022
  • Pricing Strategies, Spring 2022
  • Pricing Decision Making & Implementation, Spring 2022
  • Marketing Strategy, Fall 2022