Morrison Center for Marketing and Data Analytics

Abstract: What determines the entertainment value of a game? Ely, Frankel and Kamenica (2015) develop a formal model of suspense and surprise for analyzing entertainment value. We plan to apply their methodology using data from the video game League of Legends, a popular online game that combines real-time strategy, role-playing game elements and team coordination. League of Legends presents an ideal context to study the effects of suspense and surprise on enjoyment and engagement in leisure activities. In each of League of Legends’s matches the two potential outcomes are that one team (say, the Blue team) wins or loses. A period in a match has more suspense if the variance of the next period's probability that Blue Team wins is greater. A period has more surprise if the probability that Blue team wins is further from the last period's probability. Unlike typical sporting events, the people engaged in the game are predominantly players, not fans, and they exhibit measurable behavior. This allows us to link suspense and surprise of a match to an immediate objective behavior, such as player continuation. In addition, self-reported responses to “fun” surveys are recorded and constitute a way to measure enjoyment. In order to perform this analysis, we have acquired a unique dataset containing nearly 200 variables from Riot Games Inc., characterizing the minute-by-minute state (position, gold, experience, etc.) of approximately 10 million games. Our data also indicate whether players played prior to the observed game and continued with additional games afterward. For a small subset of the data, we also observe surveyed enjoyment.

Abstract: We study the relationship between online reviews and advertising spending in the hotel industry. Combining a dataset of TripAdvisor reviews with other datasets describing these hotels’ advertising expenditures, we show, first, that online ratings have a causal demand-side effect on ad spending. Second, this effect is negative: Hotels with higher ratings spend less on advertising than hotels with lower ratings. This suggests that hotels treat TripAdvisor ratings and advertising spending as substitutes, not complements. Third, the relationship is stronger for independent hotels than for chains, and stronger in less differentiated markets than in more differentiated markets. The former suggests that a strong brand name continues to provide some immunity to reviews, and the latter suggests that the advertising response is stronger when ratings are more likely to be pivotal. Finally, we show that the relationship between online ratings and advertising has strengthened over time, just as TripAdvisor has become more popular, implying that firms respond to online reviews if and only if consumers respond to them.

Abstract: In 2012, the state of Washington created a legal framework for production and retail sales of marijuana. Ten other U.S. states and Canada have followed. These states hope to generate tax revenue for their state budgets while limiting harms associated with marijuana sales and consumption. We use a unique administrative dataset containing all transactions in the history of the industry in Washington to evaluate the effectiveness of different tax and regulatory policies under consideration by policymakers and study the role of imperfect competition in determining these results. We use both a reduced form sufficient statistic approach and structural methods to show a number of results. First, Washington’s strict cap on firm entry has resulted in retailers with substantial market power. This market power has immediate consequences for both state tax revenue and consumer welfare. Second, because these entry restrictions have caused retailers to behave like local monopolists, the state could substantially increase revenue generated from marijuana legalization by acting as the retailer itself, as it did for alcohol sales until 2012, without a large increase in prices. Third, despite having the nation’s highest tax rate at 37%, marijuana in Washington is not overtaxed, as many policymakers in other states have argued. The high taxes do not result in lower revenue or a substantial black market. Instead, Washington is still on the upward-sloping portion of the Laffer curve, and the amount of revenue generated by a tax increase is significantly larger due to retailer market power than it would be under perfect competition. Our results suggest there is not widely available black market marijuana competing with legal retail sales. Finally, the high excise tax is primarily borne by consumers and not by firms, and there is a large social cost associated with each dollar raised.

Abstract: We study the market for fake product reviews on Amazon.com. These reviews are purchased in large private internet groups on Facebook and other sites. We handcollect data on these markets to characterise the types of products that buy fake reviews and then collect large amounts of data on the ratings and reviews posted on Amazon for these products, as well as their sales rank, advertising and pricing behavior. We use this data to assess the costs and benefits of fake reviews to sellers and evaluate the degree to which they harm consumers. The theoretical literature on review fraud shows there exist conditions when they harm consumers and conditions where they function as simply another type of advertising. Using detailed data on product outcomes before and after they buy fake reviews we can directly determine if these are low-quality products using fake reviews to deceive and harm consumers or if they are possibly high-quality products who solicit reviews to establish reputations. We find that a wide array of products purchase fake reviews including products with many reviews and high average ratings. Soliciting fake reviews on Facebook leads to a significant increase in average rating and sales rank but the effect disappears after roughly 1 month. After firms stop buying fake reviews their average ratings fall significantly and the share of one-star reviews increases significantly, indicating fake reviews are mostly used by low quality products and are deceiving and harming consumers. We also observe that Amazon deletes large numbers of reviews and we document their deletion policy.

Abstract: CEOs and companies are increasingly speaking out about controversial issues such as abortion, gun control, transgender rights, and race. There is little research, however, on whether corporate political activism affects political participation and interest in political topics. The goal of this study is to assess whether CEOs publicly taking a position on a controversial issue affects rates of political participation or interest in politics, and whether these effects differ for Democrats and Republicans.

Abstract: How do consumers value sets of alternatives and choose among them? The normative value of a set is derived from the expected value of the best option in the set. As a result, adding options to a set can increase the normative value but not decrease it. Yet prior work has found that the perceived value of a choice set systematically deviates from this normative benchmark: When less-attractive options are included in a set, the perceived value of the set decreases. In this research, we study how such undervaluation relates to visual attention by measuring what consumers attend to during choice. Prior research has found a causal role of attention in choices between liked items: The more attention an option receives, the more likely it is to be chosen. In this work, we examine how allocation of attention among options within a set relates to the ultimate choice of that set and implied perceived value. Through measuring visual attention, we hope to better understand how value is integrated across alternatives when one immediate choice opens the door to additional choices down the road.

Abstract: We study how promotions affect consumer behavior on an online retailing platform in the long term. We focus on a specific promotion: offering consumers coupons for products that have been in their shopping carts for more than one day. In a randomized field experiment involving more than 100 million customers with Alibaba Group—China’s largest e-commerce company—we randomly assigned half of eligible customers to a treatment condition where they might receive promotions for products in their shopping carts, while the other half of eligible customers did not receive coupons. We document unintended consequences of this promotion program during the month following our treatment period. On the positive side, our promotion program boosted consumer engagement, increasing the daily number of products customers viewed as well as customers’ daily purchase likelihood on the platform during the post-treatment period. On the negative side, we find that our promotion program intensified strategic consumer behaviors in two ways. First, receiving our promotions in the treatment period increased the proportion of products that consumers added to their shopping carts upon viewing them in the post-treatment period, possibly due to customers’ anticipation of promotions targeted at products in their shopping carts. Second, receiving our promotions in the treatment period lowered the prices customers paid for products in the post-treatment period. This effect holds for products that did not offer shopping-cart promotions, suggesting that prior use of shopping-cart promotions made people more price sensitive and trained them to search for other promotion mechanisms beyond shopping-cart-specific promotions. Importantly, both the positive and negative long-term effects spilled over to sellers on the same retailing platform that did not previously offer promotions to consumers. Our findings suggest that price promotions may change expectations and reference points, which can further produce both a positive long-term effect on consumer engagement and a negative long-term effect on strategic behavior. We discuss the practical implications of our findings for platforms and retailers.

Read Publication

Abstract: We study the value of short-lived and experiential-oriented pop-up stores, a popular type of omnichannel retail strategy, on both retailers that participate in pop-up store events and retailing platforms that host these retailers. We conduct a large-scale, randomized field experiment with Alibaba Group involving approximately 800,000 customers. We randomly assign consumers to either receive a message about an upcoming weeklong pop-up store event organized by Alibaba’s business-to-consumer platform (Tmall.com) or not receive any message about the event. We find that our message increased foot traffic to the pop-up store and in turn boosted expenditure at participating retailers’ online stores at Tmall after the event ended. Furthermore, we use advanced Wi-Fi technology to track customers’ visits to the pop-up store—a missing component from past research that commonly relies on point-of-sales data. We find that pop-up store visits substantially increased customers’ subsequent expenditure at participating retailers’ Tmall stores. In addition, from a platform perspective, we show that pop-up store visits increased customers’ purchases at retailers that sell related products on Tmall but did not participate in the pop-up store event. Additional analyses shed light on possible mechanisms underlying the cross-channel and spillover effects of pop-up stores and demonstrate that these effects were concentrated on prospective consumers.

Abstract: I develop and test a theory of goal maintenance that posits that individuals who achieve a life-changing goal — such as getting out of debt, becoming sober or losing a substantial amount of weight — are more likely to maintain the progress achieved during goal pursuit if they psychologically distance themselves from the pre-goal self and routinely engage in activities that activate memories of the past, less-flattering self. This theory of goal maintenance builds on prior research in identity appraisal (Wilson & Ross, 2001), vividness and intertemporal choice (Hershfield et al., 2011), and self-discrepancy (Higgins, 1987) as it relates intertemporal discrepancies in self-state representation to motivation and behavior. I applied this theory of goal maintenance to weight-loss maintenance. Through a series of six studies, I provide evidence that goal maintenance is a distinct psychological phenomenon from goal pursuit along the dimensions of past self-salience and psychological distance; and I also show that activating memories of a past, overweight self and feeling more psychologically distant from this self lead to implicit goal maintenance behavior, such as a higher willingness to pay for healthy versus unhealthy items and greater interest in learning about healthy behaviors and topics. Additionally, I provide evidence across studies that past self-salience is more associated with a prevention regulatory focus (i.e., preventing unhealthy behaviors), whereas psychological distance is more associated with a promotion regulatory focus (i.e., promoting healthy behaviors). Further, a longitudinal study of a small sample of individuals examined whether the positive effects of salience and psychological distance on weight-maintenance behaviors may persist over time and outside of a laboratory environment.

Read Publication