Time: 10:10 - 11:30

Chair: Takashi Teramoto (Chuo University)

Kohei Hayashida (U of California San Diego / University College London)

Title: The Economic Value of Algorithmic Match Recommendations: Evidence from a Natural Experiment on a Dating Platform

Abstract:  Two-sided matching platforms rely on algorithmic recommenders to deliver cross-side exposure-the stream of opposite-side users who appear in a user's inbox-yet causal evidence on whether this exposure drives subscription and retention, rather than matching efficiency alone, is scarce. We exploit a natural experiment at a large Japanese online dating platform in which a temporary, unintended interruption in recommendation exposure left newly registered users temporarily absent from other users' recommendation feeds, so they received no incoming likes. Browsing, outbound likes, and the resulting match notifications were unaffected-a clean shock to whether users were recomnended at all. Using an hourly panel of newly registered users tracked over five calendar weeks spanning the interruption, we find economically and statistically significant declines in subscription and increases in account deletion among affected users-the subscription decline concentrated among men, and account deletion rising for both men and women-with effects that are immediate and persist through the first week. The loss operates predominantly through the missing match-the interruption shifts users into a no-match state where subscription is far lower and deletion far higher. Cross-side exposure is thus a monetization and retention asset, underscoring the returns to recomnender coverage and reliability.

Keywords: Two-sided matching markets, recommendation systems, online dating, platform design

Peng Liu (Santa Clara University)  

Title: Why High-Quality Restaurants Deploy More Fake Reviews: Quality Alignment and Manipulation Persistence

Abstract: Whether strategic review manipulation distorts market information or merely amplifies existing quality signals is an open question with direct implications for platform design. Using ground-truth labeled data from 1,782 restaurants on China's largest review platform, we show that high-quality restaurants deploy fake reviews at significantly higher rates than low-quality restaurants, particularly under competitive pressure, a pattern that challenges the reputation-repair assumption in prior work. We develop a twostage dynamic model estimated via Bayesian MCMC to identify the mechanism driving this asymmetry. Fake reviews generate persistent market benefits only when consistent with underlying quality: high quality restaurants retain 68% of the manipulation benefits after one week, while low-quality restaurants retain only 5%. Low-quality restaurants also bear disproportionate long-run costs: they retain 65% of the negative discrepancy effects, compared with 11% for high-quality restaurants, at an estimated 42,649 Yuan annually per 0.05 rating discrepancy, more than twice the 17,371 Yuan that high-quality restaurants face. These findings reveal that review systems exhibit partial self-correcting properties: manipulation amplifies rather than distorts quality differentiation, with low-quality manipulation largely self-limiting and high-quality manipulation persisting because subsequent consumer experiences are consistent with the elevated expectations the fake reviews create. For managers, fake reviews pay off only when authentic quality can sustain the expectations they create. For platforms, natural market feedback disciplines low-quality manipulation effectively but underestimates the persistence of qualityaligned strategic reputation management.

Keywords: Digital platforms, Online reviews, Review manipulation, Quality signaling, Dynamic modeling.

Paper: URL

Jenny (Jiyeon) Lee (U of New South Wales)

Title: Credit-financed Experiential Consumption Under Inflation

Abstract: Research on how consumers finance experiential purchases during sustained inflation remains limited. Drawing on intertemporal consumption theory, this study investigates how inflation expectations influence the relationship between consumer credit use and experiential consumption, and whether this effect differs across levels of financial numeracy. Using multiple U.S. datasets from 2013 to 2023, we find that consumers rely on available credit to bring forward experiential spending in anticipation of rising prices. We also observe variation in this relationship based on financial numeracy. This study offers empirical insights into how inflation expectations shape intertemporal consumption behavior and provides practical implications for firms and policymakers in inflationary environments.

Keywords: Inflation Expectations, Credit, Experiential Consumption

Akira Shimizu (Keio University)

Title: Differences in Factors Influencing Alcoholic Beverage Choices in Australia and Japan

Abstract:  In this study, we investigate differences in the factors influencing wine and beer choice between Australia and Japan. Although prior research has shown that cultural factors significantly influence the choice of alcoholic beverages, these factors are measured arbitrarily. We identified cultural variables from consumer behavior theory (Study 1) and used them as control variables to examine differences in wine and beer choices between the two countries (Study 2). Subsequently, we examined differences in the meaning of the “green scale,” which has been identified as important in wine choice, between the two countries (Study 3). In accordance with global marketing research, we conducted a reliability analysis (Study 1) and a measurement invariance assessment (Study 2) to verify the validity of the variables. Study 1 revealed that cultural differences between the two countries can be explained by several variables. In Study 2, for wine, the green scale was a significant factor in Australia, but not in Japan. Study 3 identified that the green scale has different meanings in the two countries. These findings demonstrate that cultural factors are explained by consumer behavior theories, and considering both the validity of the variables and differences in their interpretation is important.

Keywords: Culture difference, Global comparison, Alcoholic beverages

Time: 11:14 - 13:00

Chair:  Kyung Tae Lee (Chuo University)

Shosuke Noguchi (Waseda University)

Title: Strategic Delay and Herding under Inventory Scarcity: Evidence from Flower Auctions

Abstract: In the presence of inventory availability information and a declining price path, buyers face a dynamic trade-off: strategically delaying their purchase to secure a lower price versus acting immediately to avoid a stockout. We examine this tension within the context of the multi-unit descending (Dutch) auction, utilizing data from the flower wholesale market. We document empirical evidence of both strategic delay under abundant inventory and scarcity-driven herding once prior purchases deplete the stock. To quantify these dynamics, we develop and estimate a structural model of buyer valuations and dynamic decision-making. Our counterfactual simulations reveal that while the sequential Dutch auction yields slightly lower expected revenue than efficient alternative auction formats, it compensates by providing superior revenue stability and faster sales velocity. Furthermore, it generates significantly higher revenue and substantially lower waste rates than the posted-price  mechanism, even if sellers can observe the majority of buyers' valuations. Ultimately, we demonstrate that the sequential Dutch auction serves as an optimal mechanism for risk-averse sellers processing multi-unit perishable goods within a strict timeline.

Keywords: Forward-looking behavior, Limited availability, Delay, Herding, Dynamic auction.

Minjung Kwon (Syracuse University)

Title: Prescribing Sustainability: How ESG Shapes Physicians’ Drug Choices and Pharmaceutical Pricing? 

Abstract:  Problem definition: This study examines how ESG incidents propagate through pharmaceutical supply chains by shaping downstream physicians’ prescribing decisions and firms’ pricing strategies. Methodology/results: Using granular claim-level prescription data, we show that physicians prescribe fewer brand-name drugs from firms implicated in negative ESG incidents, accompanied by a notable decline in unit prices. Controlling for physician- and product-level heterogeneity, we estimate a 9% reduction in daily dispensed quantities and a 21% decrease in unit prices at the time of ESG incidents. These effects are absent for generic drugs, for which physicians rarely specify the manufacturer. The decline in brand-name prescriptions appears to be driven by heightened product-quality concerns and is more pronounced in counties with higher ESG awareness—proxied by higher levels of education, income, and Democratic voting shares. Share-of-Wallet analyses among close substitutes provide direct evidence that physicians shift away from affected firms, reducing brand stickiness. Managerial implications: This study identifies physicians as an important but overlooked ESG-responsive stakeholder group and extends the literature on sustainability, corporate social irresponsibility, and healthcare supply chains. By linking upstream ESG incidents to downstream prescribing decisions and firms’ pricing responses, our findings show that broader ESG externalities, beyond functional product failures, can fundamentally reshape professional decision-making and market outcomes, even in healthcare settings where clinical utility is the overriding mandate. Managers and decision makers should recognize that ESG failures can affect clinical-market demand, brand loyalty, and pricing power, and supply chain resilience; pharmaceutical firms therefore need to treat ESG performance as a strategic capability, not merely a compliance or reputational concern.

Keywords: ESG, Pharmaceutical Supply Chains, Prescription, Physician Behavior, Strategic Pricing, Brand Switching.

Richard Tang (LMU), Desmond Lo (Santa Clara University), Satoshi Fukuda (SCU), Chieko Minami (Tokyo Woman’s Christian University)

Title: Weather Volatility, Organization Design, and Retail Performance

Abstract: Weather volatility increasingly disrupts retail businesses. Yet we know relatively little about how firms should organize internally to adapt to such shocks. This paper examines when the delegation of decision rights helps or hinders retail performance amid weather volatility. We argue that delegation has a conditional effect on sales performance. Delegating authority to department managers enables fast, locally informed responses to weather shocks, but it can also weaken coordination across departments. Thus, delegation improves performance when local adaptation is valuable and coordination needs are limited, but it reduces performance when cross-departmental alignment is central.
Based on an analytical model of information acquisition, we test this argument using a rich panel from a major Japanese general-merchandise retailer. The data combine high-resolution rainfall observations from the Japan Meteorological Agency, an original survey of department managers on delegation and coordination, and proprietary monthly department-level performance records. The setting covers 350 stores, 24 department types, and about 2700 managers. We measure weather volatility by a store-specific, calendar-month measure of upper-tail rainfall shocks relative to each store’s own historical rainfall distribution, while performance is measured by the ratio of realized monthly sales to assigned monthly quota.
The empirical results show that the value of delegation depends strongly on coordination conditions. For perishable-product departments, such as deli, fish, processed meat, produce, and daily food, delegation under rainfall volatility improves performance. These departments face short shelf lives and steep penalties due to slow local adjustment. In contrast, for departments with high coordination importance or dense peer-manager networks, delegation under volatility reduces performance, consistent with the cost of uncoordinated local action. The pattern holds for both weather-specific task delegation and broader sales-task delegation, as well as alternative measures of coordination importance.
The paper contributes to marketing and organizational design research by underscoring that firm adaptation to climate-related volatility is not only a matter of forecasting but also depends on who holds decision-making authority within the firm and on whether the task environment rewards local speed or coordinated alignment.

Keywords: Weather volatility, delegation of decision rights, sales and retail management, organizational design, cross-unit coordination.

Yutec Sun (French National Graduate School of Statistics and Information Analysis/Korea Development Institute)

Title: Competition and Quality: Evidence from the French Mobile Network Services

Abstract: Across service sectors, the quality of service is produced using congestible infrastructure. We study dynamic network investment competition among mobile services in France to investigate how competition shapes infrastructure allocation, quality, and welfare under production congestion. We estimate a congestible quality production technology, consumer demand, service supply, and investment costs to characterize long-run equilibrium quality provision and its welfare implications under competition, monopoly, and the social planner's allocation. We find that input misallocation does not track distortions in service quality under production congestion: less productive firms overinvest in infrastructure yet undersupply quality because congestion from their excess subscribers offsets investment gains, while a monopolist underinvests yet oversupplies quality by restricting demand and reducing congestion. Monopoly misallocation generates welfare losses 4.7–6.4 times larger than under competition. Entry reduces incumbents' investment yet improves their quality through congestion relief, and is unambiguously welfare-enhancing.

Keywords: Congestion, Misallocation, Quality, Competition, Investment

Time: 14:30 - 15:50

Chair: Makoto Abe (Chuo University)

Jangwon Choi (Santa Clara University)

Title: A Flexible Framework for Ad Stock Estimation: A Bayesian Alternative to Geometric Decay

Abstract: One of the earliest and most extensive literatures in quantitative marketing concerns the measurement of ad effectiveness. Because ads do not always “work” immediately, and often require multiple exposures, econometric approaches to measuring the cumulative impact of advertising typically rely on the concept of a (latent) “ad stock” or “goodwill.” Owing to the influential work of Nerlove and Arrow (1962), it has often been assumed that the contribution of each ad to the ad stock decays exponentially over time at a constant rate, captured by a single parameter common across ad types. In this project, we propose Bayesian alternatives to this restrictive specification across two distinct empirical analyses. First, we examine how two different types of online advertising campaigns, acquisition and retargeting, differentially affect online users’ behaviors. By proposing a flexible, semiparametric Bayesian approach to regularize past ad weights, we analyze an online panel of individual-level ad impression data for a French financial services firm. We find empirical evidence that the traditional single-parameter decay assumption fails to accurately capture the differing effects of acquisition and retargeting ads on users’ propensities to visit the target firm’s website. In the second analysis, we apply Gaussian Process (GP) priors, a Bayesian nonparametric approach to estimating functions, to flexibly account for the decaying impact of past ad spendings by major cell phone manufacturers in the US market. Together, these applications demonstrate that flexible Bayesian models are able to provide a more accurate and nuanced understanding of cumulative advertising dynamics than traditional exponential decay frameworks.

Keywords: Ad stock, Nerlove-Arrow, Bayesian semiparametric models, Gaussian processes.

Masakazu Ishihara (New York University) Baek Jung Kim (Korea University) & Hiroshi Kumakura (Chuo University)

Title: The Role of Time-shifting in Serial Drama Viewing: Implications for Drama Producers and Advertisers

Abstract: Despite its important implications for TV program producers and advertisers, little has been investigated regarding consumers' choice of TV viewing methods- live versus time-shifting (i.e., viewing with DVRs). This paper studies how consumers watch TV programs and the role of time-shifting by using unique individual-level TV viewing data in Japan. We develop a dynamic structural model of consumers' TV viewing and ad-exposure decisions. Using the parameter estimates, we conduct a counterfactual experiment and quantify the value of time-shifting (by removing the time-shifting option). Consumers might skip ads more while time-shifting, but it can increase the number of drama viewers directly (e.g., consumers on a night-shift job cannot watch prime-time dramas without DVRs) or indirectly (e.g., those who missed an episode might stop watching thereafter). The results show that if time-shifting is not available, (1) the average total audience rating decreases by 7.9 percentage points (from 21.7% to 12.8%), and (2) the aggregated total ad exposure decreases by 27.6 percentage. We also explore how these results vary across dramas and find that this pattern seems more substantial for unpopular dramas (rather than popular ones). Thus, time-shifting for serial dramas in Japan helps both TV producers and advertisers, and particularly, unpopular dramas benefit more.

Keywords: TV Viewing, Time-shift, Product Completeness, Dynamic Programming

Maggie Dong (U of New South Wales)

Title: Content Release Strategy and User Engagement on a Major AVoD Platform

Abstract: Streaming platforms increasingly differ in how they release serialized content—either all at once (“binge”) or sequentially (“episodic”)—yet little is known about how these strategies shape user engagement in advertising-based video-on-demand (AVoD) environments. Using three years of viewing records from a major AVoD platform, we examine how release strategy affects both within-show engagement and cross-show spillover. Binge-released shows yield about a third fewer chapter views than its non-binge twin, but more catalog breath and deeper completion. These effects are robust across propensity-score matching and instrumental-variable specifications and vary with show characteristics such as genre and narrative structure. The cliffhanger strength of the content moderates the effects, with a strong cliffhanger lifts non-binge viewers' engagement but not binge viewers. The findings highlight tradoffs for release strategy , offering guidance for AVoD platforms and content producers.

Keywords: Video-on-Demand (VoD), Content Release Strategy, Streaming Platforms, Advertising-based Video-on-Demand (AVoD).

Bryan K. Bollinger  (Dartmouth University), Kenneth Gillingham (Yale University) & A. Justin Kirkpatrick (Michigan State University)

Title: Self-and Social Signaling: Evidence from Solar Adoption in California

Abstract: One of the earliest and most extensive literatures in quantitative marketing concerns the measurement of ad effectiveness. Because ads do not always “work” immediately, and often require multiple exposures, econometric approaches to measuring the cumulative impact of advertising typically rely on the concept of a (latent) “ad stock” or “goodwill.” Owing to the influential work of Nerlove and Arrow (1962), it has often been assumed that the contribution of each ad to the ad stock decays exponentially over time at a constant rate, captured by a single parameter common across ad types. In this project, we propose Bayesian alternatives to this restrictive specification across two distinct empirical analyses. First, we examine how two different types of online advertising campaigns, acquisition and retargeting, differentially affect online users’ behaviors. By proposing a flexible, semiparametric Bayesian approach to regularize past ad weights, we analyze an online panel of individual-level ad impression data for a French financial services firm. We find empirical evidence that the traditional single-parameter decay assumption fails to accurately capture the differing effects of acquisition and retargeting ads on users’ propensities to visit the target firm’s website. In the second analysis, we apply Gaussian Process (GP) priors, a Bayesian nonparametric approach to estimating functions, to flexibly account for the decaying impact of past ad spendings by major cell phone manufacturers in the US market. Together, these applications demonstrate that flexible Bayesian models are able to provide a more accurate and nuanced understanding of cumulative advertising dynamics than traditional exponential decay frameworks.

Paper: URL

Time: 16:00 - 16:45

Chair: Makasazu Ishihara (New York University)

Bryan K. Bollinger  (Dartmouth University)

Title: The Frontier of Quant Marketing Research & Some Tips for the Publication Process by co-Editor of JMR (Meet-the-Editor)