Xin (Shane) Wang

• Wang, X.; Ryoo, J.; Bendle, N. T.; Kopalle, P. K., 2021, "The role of machine learning analytics and metrics in retailing research", Journal of Retailing, December 97(4): 658 - 675.
  Abstract: This research presents the use of machine learning analytics and metrics in the retailing context. We first discuss what is machine learning and explain the field’s origins. We then demonstrate the strengths of machine learning methods using an online retailing dataset, noting key areas of divergence from the traditional explanatory approach to data analysis. We then provide a review of the current state of machine learning in top-level retailing and marketing research, integrating ideas for future research and showcasing potential applications for practitioners. We propose that the explanatory and machine learning approaches need not be mutually exclusive. Particularly, we discuss four key areas in the general scientific research process that can benefit from machine learning: data exploration/theory building, variable creation, estimation, and predicting an outcome metric. Due to the customer-facing nature of retailing, we anticipate several challenges researchers and practitioners might face in the adoption and implementation of machine learning, such as ethical prediction and customer privacy issues. Overall, our belief is that machine learning can enhance customer experience and, accordingly, we advance opportunities for future research.

  Link(s) to publication:

  http://dx.doi.org/10.1016/j.jretai.2020.12.001

• Wang, X.; Lu, S.; Li, X. I.; Khamitov, M.; Bendle, N. T., 2021, "Audio Mining: The Role of Vocal Tone in Persuasion", Journal Of Consumer Research, August 48(2): 189 - 211.
  Abstract: Persuasion success is often related to hard-to-measure characteristics, such as the way the persuader speaks. To examine how vocal tones impact persuasion in an online appeal, this research measures persuaders’ vocal tones in Kickstarter video pitches using novel audio mining technology. Connecting vocal tone dimensions with real-world funding outcomes offers insight into the impact of vocal tones on receivers’ actions. The core hypothesis of this paper is that a successful persuasion attempt is associated with vocal tones denoting (1) focus, (2) low stress, and (3) stable emotions. These three vocal tone dimensions—which are in line with the stereotype content model—matter because they allow receivers to make inferences about a persuader’s competence. The hypotheses are tested with a large-scale empirical study using Kickstarter data, which is then replicated in a different category. In addition, two controlled experiments provide evidence that perceptions of competence mediate the impact of the three vocal tones on persuasion attempt success. The results identify key indicators of persuasion attempt success and suggest a greater role for audio mining in academic consumer research.

  Link(s) to publication:

  http://dx.doi.org/10.1093/jcr/ucab012

• Ryoo, J.; Wang, X.; Lu, S., 2021, "Do Spoilers Really Spoil? Using Topic Modeling to Measure the Effect of Spoiling Movie Reviews on Box Office Revenue", Journal of Marketing, March 85(2): 70 - 88.
  Abstract: A sizable portion of online movie reviews contain spoilers, defined as information that prematurely resolves plot uncertainty. In this research, the authors study the consequences of spoiler reviews using data on box office revenue and online word of mouth for movies released in the United States. To capture the degree of information in spoiler review text that reduces plot uncertainty, the authors propose a spoiler intensity metric and measure it using a correlated topic model. Using a dynamic panel model with movie fixed effects and instrumental variables, the authors find a significant and positive relationship between spoiler intensity and box office revenue with an elasticity of .06. The positive effect of spoiler intensity is greater for movies with a limited release, smaller advertising spending, and moderate user ratings, and is stronger in the earlier days after the movie’s release. Using an event study and online experiments, the authors provide further evidence that spoiler reviews can help consumers reduce their uncertainty about the quality of movies, consequently encouraging theater visits. Thus, movie studios may benefit from consumers’ access to plot-intense reviews and should actively monitor the content of spoiler reviews to better forecast box office performance.

  Link(s) to publication:

  http://dx.doi.org/10.1177/0022242920937703

• Nguyen, P.; Wang, X.; Li, X.; Cotte, J., 2021, "Reviewing Experts’ Restraint from Extremes and its Impact on Service Providers", Journal of Consumer Research, February 47(5): 654 - 674.
  Abstract: This research investigates reviewing experts on online review platforms. The main hypothesis is that greater expertise in generating reviews leads to greater restraint from extreme summary evaluations. The authors argue that greater experience generating reviews facilitates processing and elaboration, and enhances the number of attributes implicitly considered in evaluations, which reduces the likelihood of assigning extreme summary ratings. This restraint-of-expertise hypothesis is tested across different review platforms (TripAdvisor, Qunar, and Yelp), shown for both assigned ratings and review text sentiment, and demonstrated both between (experts vs. novices) and within reviewers (expert vs. pre-expert). Two experiments replicate the main effect and provide support for the attributes-based explanation. Field studies demonstrate two major consequences of the restraint-of-expertise effect. (i) Reviewing experts (vs. novices), as a whole, have less impact on the aggregate valence metric, which is known to affect page-rank and consumer consideration. (ii) Experts systematically benefit and harm service providers with their ratings. For service providers that generally provide mediocre (excellent) experiences, reviewing experts assign significantly higher (lower) ratings than novices. This research provides important caveats to the existing marketing practice of service providers incentivizing reviewing experts, and provides strategic implications for how platforms should adopt rating scales and aggregate ratings.

  Link(s) to publication:

  http://dx.doi.org/10.1093/jcr/ucaa037

• Vakratsas, D.; Wang, X., 2021, "Artificial Intelligence in Advertising Creativity", JOURNAL OF ADVERTISING, January 50(1): 39 - 51.
• Lu, S.; Wang, X.; Bendle, N. T., 2020, "Does Piracy Create Online Word-of-Mouth? An Empirical Analysis in Movie Industry", Management Science, May 66(5): 2140 - 2162.
  Abstract: Anecdotal evidence suggests that counterfeiting/piracy can help create online word-of-mouth (WOM) and through this boost demand but how powerful is such WOM? To answer this question, we conduct a descriptive study with some attempts to establish near causality. We estimate the impact of piracy on WOM and ultimately revenue by applying a panel data method to all movies widely released in the U.S. from 2015 to 2017. In identifying the effects of piracy we make inventive use of Russian piracy data to construct instrument variables for piracy in the U.S. This is possible as the key piracy site, the Pirate Bay, has been blocked in Russia since 2015. We find movies with pre-release piracy are associated with lower revenues despite the WOM effect. Critically, however, we show a positive correlation between post-release piracy and WOM volume and, extend the field, by finding that the presence of post-release piracy is associated with about 3.0% increase in box office revenue. We also note the impact of a raid by the Swedish Police that temporarily took down the Pirate Bay website in December 2014. The period when the site was down experienced a decline in WOM volume and revenues, consistent with the effect of lower post-release piracy predicted by our models. Our findings suggest approaches to target scarce anti-piracy resources, such as focusing on tackling damaging pre-release piracy.

  Link(s) to publication:

  https://ssrn.com/abstract=3308905

  http://dx.doi.org/10.1287/mnsc.2019.3298

• He, J.; Wang, X.; Vandenbosch, M.; Nault, B., 2020, "Revealed Preference in Online reviews: Purchase Verification in the Tablet Market", Decision Support Systems, May 132
  Abstract: The review systems of online platforms create a stream of online word-of-mouth that allows consumers to learn from others' purchasing experience. However, it is difficult for consumers to discern the authenticity of a review or the reviewer's level of experience with the product. Platforms can aid the authentication process by incorporating a verified purchase (VP) indication, or “badge” as is done on Amazon, in reviews where the consumer writing the review has verifiably purchased the focal product. A VP is a revealed preference for a product implying a utility-maximizing choice where the consumer writing the review has experience with the product. Combining an Amazon dataset on tablet computers with the theory of revealed preference in online reviews, we uncover a surprising new result: the proportion of VP reviews (a revealed preference) is associated with higher future sales, and the effect of the proportion of VP reviews on sales dominates the effect of the mean rating. This novel use of VP with revealed preference theory has implications for new research in the design of recommendation systems, detecting fraudulent reviews, and online profiling/privacy. Moreover, the use of a VP badge is immediately applicable to firms and platforms.

  Link(s) to publication:

  http://dx.doi.org/10.1016/j.dss.2020.113281

• Li, X.; Li, K.; Wang, X., 2020, "Transparency of Behavior-Based Pricing", Journal of Marketing Research, February 57(1): 79 - 99.
  Abstract: Behavior-based pricing (BBP) refers to the practice in which firms collect consumers’ purchase history data, recognize repeat and new consumers from the data, and offer them different prices. This is a prevalent practice for firms and a worldwide concern for consumers. Extant research has examined BBP under the assumption that consumers observe firms’ practice of BBP. However, consumers do not know that specific firms are doing this and are often unaware of how firms collect and use their data. In this article, the authors examine (1) how firms make BBP decisions when consumers do not observe whether firms perform BBP and (2) how the transparency of firms’ BBP practice affects firms and consumers. They find that when consumers do not observe firms’ practice of BBP and the cost of implementing BBP is low, a firm indeed practices BBP, even though BBP is a dominated strategy when consumers observe it. When the cost is moderate, the firm does not use BBP; however, it must distort its first-period price downward to signal and convince consumers of its choice. A high cost of implementing BBP serves as a commitment device that the firm will forfeit BBP, thereby improving firm profit. By comparing regimes in which consumers do and do not observe a firm’s practice of BBP, the authors find that transparency of BBP increases firm profit but decreases consumer surplus and social welfare. Therefore, requiring firms to disclose collection and usage of consumer data could hurt consumers and lead to unintended consequences.

  Link(s) to publication:

  http://dx.doi.org/10.1177/0022243719881448

• Khamitov, M.; Wang, X.; Thomson, M., 2019, "How Well Do Brand Relationships Drive Customer Brand Loyalty? Generalizations from a Meta-Analysis of Brand Relationship Elasticities", Journal of Consumer Research, October 46(3): 435 - 459.
  Abstract: To advance understanding of how well different types of brand relationships drive customer brand loyalty and to help companies improve the effectiveness of their relationship-building investments, this article conducts a meta-analysis of the link between five consumer-brand relationship constructs and customer brand loyalty. The analysis of 588 elasticities from 290 studies reported in 255 publications over 24 years (n = 348,541 across 46 countries) reveals that the aggregate brand relationship elasticity is .439. More importantly, results demonstrate under what conditions various types of brand relationships increase loyalty. For example, while elasticities are generally highest for love-based and attachment-based brand relationships, the positive influence of brand relationships on customer brand loyalty is stronger in more recent (vs. earlier) years, for non-status (vs. status) and publicly (vs. privately) consumed brands as well as for estimates using attitudinal (vs. behavioral) customer brand loyalty. Overall, the results suggest that brand relationship elasticities vary considerably across brand, loyalty, time, and consumer characteristics. Drawing on these findings, the current research advances implications for managers and scholars and provide avenues for future research.

  Link(s) to publication:

  http://dx.doi.org/10.1093/jcr/ucz006

• Xi, L.; Shi, M.; Wang, X., 2019, "Video Mining: Measuring Visual Information Using Automatic Methods", International Journal of Research in Marketing, June 36(2): 216 - 231.
• Borah, A.; Wang, X.; Ryoo, J., 2018, "Understanding Influence of Marketing on Business Practice: Analysis of Business Journals Using Textual Information and Latent Dirichlet Allocation Analysis", Customer Needs and Solutions, December 5(3-4): 146 - 161.
  Abstract: Several calls have been made to understand the influence of marketing thought on practice (Rust et al. J Mark 68:76–89, 11). Practice includes practitioners who mostly use concepts and frameworks (general practice) and who mostly use quantitative models (quantitative practice). This paper compares the relative influence of marketing thought compared to other disciplines and uncovers seminal marketing thoughts that have influenced both general and quantitative practice. Using topic modeling procedures on 94 years of Harvard Business Review, 46 years of Sloan Management Review, and 47 years of Management Science, this paper illuminates the evolution of the influence of marketing thought over time. Despite marketing’s slow start, it has an increasing influence on both general and quantitative practice. Foundational topics in marketing such as product, promotion, place, consumers, and marketing research methods have influenced both general and quantitative practice. Surprisingly, price has not influenced practice. Marketing Communications is increasingly influential while Channel Management, Product/Service Management, and surprisingly Customer Relationships have lost their early influence to practice. General practitioners find Marketing Environment and Business Models increasingly influential while quantitative practitioners find Social Influence and Metrics increasingly influential. Quantitative practice has kept up to speed with marketing thought that influence general practice.

  Link(s) to publication:

  http://dx.doi.org/10.1007/s40547-018-0089-z

• Mai, F.; Shan, J.; Bai, Q.; Wang, X.; Chiang, R. H. L., 2018, "How Does Social Media Impact Bitcoin Value? A Test of the Silent Majority Hypothesis", Journal of Management Information Systems, March 35(1): 19 - 52.
  Abstract: As the world's first completely decentralized digital payment system, the emergence of Bitcoin represents a revolutionary phenomenon in financial markets. This study examines the dynamic relationships between social media and bitcoin performance. We consider the distinct effects of different social media platforms and different user groups subdivided by posting volume. The results suggest that more bullish forum posts have a positive effect on bitcoin returns, and the effect is stronger when we only include the posts by users who are less likely to contribute. In addition, messages on Internet forum have stronger impacts on future bitcoin market measures at a daily frequency, but microblogs’ effects are more significant at an hourly frequency.

  Link(s) to publication:

  https://dx.doi.org/10.2139/ssrn.2545957

• Aribarg, A.; Otter, T.; Zantedeschi, D.; Allenby, G. M.; Bentley, T.; Curry, D.; Dotson, M.; Henderson, T.; Honka, E.; Kohli, R., et al., 2018, "Advancing Non-Compensatory Choice Models in Marketing", Customer Needs and Solutions, March 5(1-2): 82 - 92.
  Abstract: The extant choice literature has proposed different non-compensatory rules as a more realistic description of consumers’ choice than a standard compensatory model. Some research has further suggested a two-stage sequential decision process of non-compensatory consideration and then compensatory choice, where the determinants of each stage may differ. Some aspects of non-compensatory choice modeling are under-studied. In this article, we hope to advance the understanding of non-compensatory choice models with the following aims: (a) providing an overview of existing representations for non-compensatory choice decisions, (b) discussing how such choice decisions can manifest from the economic search theoretical perspective, (c) exploring the empirical identification of non-compensatory decisions using different data, and (d) presenting applications of non-compensatory choice models in novel domains.

  Link(s) to publication:

  https://doi.org/10.1007/s40547-017-0072-0

  http://dx.doi.org/10.1007/s40547-017-0072-0

• He, J.; Wang, X.; Curry, D., 2017, "Mediation Analysis: A New Test When All or Some Variables are Categorical", International Journal of Research in Marketing, December 34(4): 780 - 798.
  Abstract: Statistical tests for mediation in consumer research typically use a regression coefficients (RegCoeff)-based framework. We present a new test based on likelihood ratio principles to complement the RegCoeff approach. We compare the new test's performance to conventional methods that use PROCESS, MEDIATION, and extensions thereof to categorical measures. Our tests address situations in which the assumptions of current methods tend not to hold that is when one or more variables may be categorical or relationships among constructs may be nonlinear. In such environments, results significantly favor the new test, which we call LRT for likelihood ratio test. For example, among 72 tests of power, LRT loses just 9 times to PROCESS. It wins outright in 45 cases and ties in 18 others. Differences are non-trivial they are statistically significant at the 0.001 level in 21 of the 45 wins. To assist researchers, we detail how and when the new test can best complement existing methods and offer software that performs the new test for any combination of continuous and categorical variables.

  Link(s) to publication:

  https://doi.org/10.1016/j.ijresmar.2017.08.001

• Bendle, N. T.; Wang, X., 2017, "Marketing Accounts", International Journal of Research in Marketing, September 34(3): 604 - 621.
  Abstract: Marketing actions frequently create long-term value yet this is often not recorded in financial accounts. The same records are typically used for internal reporting limiting both recognition of the value created by marketing, and accountability for the misuse of market-based assets. Creating comprehensive marketing accounts can mitigate the problems caused by financial accounting’s omission of market-based assets. We explain current accounting practice, outline the idea of marketing accounts, and contrast this with current accounting practice. Marketing accounts capture the value of market-based assets, applying accounting’s matching concept as consistently as possible to treat marketing as an investment where appropriate. These accounts are based upon expected value, and are feasible within accounting rules given they aim only to aid management, not investor, decision making. Marketing accounts vary between, but not within, firms, and are comprehensive and regular. Finally, they are controlled by marketers with assumptions and models recorded and approved by the chief marketing officer. We conclude by illustrating how to implement marketing accounts.

  Link(s) to publication:

  http://www.sciencedirect.com/science/article/pii/S0167811617300290

For more publications please see our Research Database