Knowledge that Transforms

In developing countries, mobile telecom networks have emerged as major providers of financial services, bypassing the sparse retail networks of traditional banks. We analyze a large individual-level data set of mobile money transactions in Tanzania to provide evidence of the impact of mobile money on alleviating financial exclusion in developing countries. We identify three types of transactions: (i) money transfers to others, (ii) short-distance money self-transportation, and (iii) money storage for short to medium periods of time. We utilize a natural experiment of an unanticipated increase in transaction fees to identify the demand for these transactions. Using the demand estimates, we find that the willingness to pay to avoid walking with cash an extra kilometer (short-distance self-transportation) and to avoid storing money at home (money storage) for an extra day are 1.25% and 0.8% of an average transaction, respectively, which demonstrates that mobile money ameliorates significant amounts of crime-related risk. We explore the implications of these estimates for pricing and demonstrate the profitability of incentive-compatible price discrimination based on type of service, consumer location, and distance between transaction origin and destination. We show that differential pricing based on the features of a transaction delivers a Pareto improvement. Data and the online appendix are available at https://doi.org/10.1287/mksc.2017.1027 .

In 2008, New York City mandated that all chain restaurants post calorie information on their menus. For managers of chain and standalone restaurants, as well as for policy makers, a pertinent goal might be to monitor the impact of this regulation on consumer conversations. We propose a scalable Bayesian topic model to measure and understand changes in consumer opinion about health (and other topics). We calibrate the model on 761,962 online reviews of restaurants posted over eight years. Our model allows managers to specify prior topics of interest such as “health” for a calorie posting regulation. It also allows the distribution of topic proportions within a review to be affected by its length, valence, and the experience level of its author. Using a difference-in-differences estimation approach, we isolate the potentially causal effect of the regulation on consumer opinion. Following the regulation, there was a statistically small but significant increase in the proportion of discussion of the health topic. This increase can be attributed largely to authors who did not post reviews before the regulation, suggesting that the regulation prompted several consumers to discuss health in online restaurant reviews. Data and the online appendix are available at https://doi.org/10.1287/mksc.2017.1048 .

In many industries, product design and manufacturing lead times are sufficiently long that both the quality level of a product and the amount of inventory produced must be determined before a firm knows what the actual demand will be. In this paper, we conduct a theoretical analysis of such a setting. We first consider a centralized channel and characterize the optimal decisions by establishing relationships that must hold between the elasticity of cost of quality and the elasticity of revenue and show that quality and inventory are strategic substitutes. Next, we consider a decentralized channel with a wholesale price contract, in which a manufacturer determines quality and wholesale price, while a retailer determines inventory and retail price. We find that, different from the case without endogenous inventory, product quality can be higher in a decentralized channel compared to a centralized channel, and this is because a wholesale price contract shields the manufacturer from inventory risk. For both centralized and decentralized channels, we find that as demand uncertainty increases, quality decreases, while, different from the case without endogenous quality, inventory can be U-shaped. Interestingly, to mitigate the impact of demand uncertainty on profit, quality can be a more effective lever than inventory in a centralized channel; however, in a decentralized channel, quality is less responsive and inventory is more responsive to demand uncertainty than in a centralized channel. The online appendix is available at https://doi.org/10.1287/mksc.2017.1041 .

Consumers engage in costly searches to evaluate the increasing number of product options available from online retailers. Presenting the best alternatives at the beginning reduces search costs associated with a consumer finding the right product. We use rich data on consumer click-stream behavior from a major web-based hotel comparison platform to estimate a model of search and click. We propose a method of determining the ranking of search results that maximizes consumers’ click-through rates (CTRs) based on partial information available to the platform at the time of the consumer request, its assessment of consumers’ preferences, and the expected consumer type based on request parameters from the current visit. Our method has two distinct advantages. First, we endogenize a consumer response to the ranking using search refinement tools, such as sorting and filtering of product options. Accounting for these search refinement actions is important since the ranking and consumer search actions together shape the consideration set from which clicks are made. Second, rankings are targeted to anonymous consumers by relating price sensitivity to request parameters, such as the length of stay, number of guests, and day of the week of the stay. We find that predicted CTRs under our proposed ranking are almost double those of the platform’s default ranking. Data and the online appendix are available at https://doi.org/10.1287/mksc.2017.1036 .

Although firms are leveraging weather conditions in promotions, they struggle to quantify the impact. This study exploits field experiment data on weather-based mobile promotions with over six million users. Results find that sunny and rainy weather have first-order main effects. Purchase responses to promotions are higher and faster in sunny weather relative to cloudy weather, whereas purchase responses to promotions are lower and slower in rainy weather. These findings are robust across different measures of weather changes with both backward-looking historical weather and forward-looking forecasts, as well as deviations from normal weather. Also, sunny and rainy weather have second-order interactive effects with ad copies of mobile promotions. Compared with the neutral ad copy, the prevention frame ad copy hurts the initial promotion boost induced by sunshine, but improves the initial promotion drop induced by rainfall. For marketers, these findings imply new opportunities in customer data analytics for more effective weather-based mobile targeting. Data and the online appendix are available at https://doi.org/10.1287/mksc.2017.1044 .

With the cooperation of a large mobile service provider, we conduct a novel field experiment that simultaneously randomizes the prices of two competing movie theaters using mobile coupons. Unlike studies that vary only one firm’s prices, our experiment allows us to account for competitor response. We test mobile targeting based on consumers’ real-time and historic locations, allowing us to evaluate popular mobile coupon strategies in a competitive market. The experiment reveals substantial profit gains from mobile discounts during an off-peak period. Both firms could create incremental profits by targeting their competitor’s location. However, the returns to such “geoconquesting” are reduced when the competitor also launches its own targeting campaign. We combine our experimentally generated data with a demand model to analyze optimal pricing in a static Bertrand–Nash equilibrium. Interestingly, competitive responses raise the profitability of behavioral targeting where symmetric pricing incentives soften price competition. By contrast, competitive responses lower the profitability of geographic targeting, where asymmetric pricing incentives toughen price competition. If we endogenize targeting choice, both firms would choose behavioral targeting in equilibrium, even though more granular geobehavioral targeting combining both real-time and historic locations is possible. These findings demonstrate the importance of considering competitor response when piloting novel price-targeting mechanisms. Data are available at https://doi.org/10.1287/mksc.2017.1042 .

The creation and sharing of user-generated content such as product reviews has become increasingly “social,” particularly in online communities where members are connected. While some online communities have used monetary rewards to motivate product review contributions, empirical evidence regarding the effectiveness of such rewards remains limited. We examine the possible moderating effect of social connectedness (measured as the number of friends) on publicly offered monetary rewards using field data from an online review community. This community saw an (unexpected) overall decrease in total contributions after introducing monetary rewards for posting reviews. Further examination across members finds a strong moderating effect of social connectedness. Specifically, contributions from less-connected members increased by 1,400%, while contributions from more-connected members declined by 90%. To corroborate this effect, we rule out multiple alternative explanations and conduct robustness checks. Our findings suggest that token-sized monetary rewards, when offered publicly, can undermine contribution rates among the most connected community members. Data and the online appendix are available at https://doi.org/10.1287/mksc.2016.1022

Firms using online advertising regularly run experiments with multiple versions of their ads since they are uncertain about which ones are most effective. During a campaign, firms try to adapt to intermediate results of their tests, optimizing what they earn while learning about their ads. Yet how should they decide what percentage of impressions to allocate to each ad? This paper answers that question, resolving the well-known “learn-and-earn” trade-off using multi-armed bandit (MAB) methods. The online advertiser’s MAB problem, however, contains particular challenges, such as a hierarchical structure (ads within a website), attributes of actions (creative elements of an ad), and batched decisions (millions of impressions at a time), that are not fully accommodated by existing MAB methods. Our approach captures how the impact of observable ad attributes on ad effectiveness differs by website in unobserved ways, and our policy generates allocations of impressions that can be used in practice. We implemented this policy in a live field experiment delivering over 750 million ad impressions in an online display campaign with a large retail bank. Over the course of two months, our policy achieved an 8% improvement in the customer acquisition rate, relative to a control policy, without any additional costs to the bank. Beyond the actual experiment, we performed counterfactual simulations to evaluate a range of alternative model specifications and allocation rules in MAB policies. Finally, we show that customer acquisition would decrease by about 10% if the firm were to optimize click-through rates instead of conversion directly, a finding that has implications for understanding the marketing funnel. Data is available at https://doi.org/10.1287/mksc.2016.1023 .

We investigate the relationship between a firm’s use of management responses and its online reputation. We focus on the hotel industry and present several findings. First, hotels are likely to start responding following a negative shock to their ratings. Second, hotels respond to positive, negative, and neutral reviews at roughly the same rate. Third, by exploiting variation in the rate with which hotels respond on different review platforms and variation in the likelihood with which consumers are exposed to management responses, we find a 0.12-star increase in ratings and a 12% increase in review volume for responding hotels. Interestingly, when hotels start responding, they receive fewer but longer negative reviews. To explain this finding, we argue that unsatisfied consumers become less likely to leave short indefensible reviews when hotels are likely to scrutinize them. Our results highlight an interesting trade-off for managers considering responding: fewer negative ratings at the cost of longer and more detailed negative feedback. Data and the online appendix are available at https://doi.org/10.1287/mksc.2017.1043 .

Product lines are ubiquitous. For example, Marriott International manages high-end ultra-luxury hotels (e.g., Ritz-Carlton) and low-end economy hotels (e.g., Fairfield Inn). Firms often bundle core products with ancillary services (or add-ons). Interestingly, empirical observations reveal that industries with ostensibly similar characteristics (e.g., customer types, costs, competition, distribution channels, etc.) employ different bundling strategies. For example, airlines bundle high-end first class with ancillary services (e.g., breakfast, entertainment) while hotel chains bundle ancillary services (e.g., breakfast, entertainment) at the low-end. We observe, unlike hotel lines that are highly differentiated at different geographic locations, airlines suffer low core differentiation because all passengers (first-class and economy) are at the same location (i.e., same plane, weather, delays, cancellations, etc.). In general, we find product lines with low core differentiation (e.g., airlines, amusement parks) routinely bundle high-end while product lines with highly differentiated cores (e.g., hotels, restaurants) routinely bundle low-end. High-end bundling makes the high-end more attractive, increasing line differentiation (less intraline competition) while low-end bundling decreases line differentiation. Therefore, bundling allows optimal differentiation given a differentiation constraint (complex costs). Last, firms may use strategic bundling for targeting in their core products; e.g., low-end hotels bundle targeted add-ons unattractive to high-end consumers such as lower-quality breakfasts and slower Internet. Data, as supplemental material, are available at https://doi.org/10.1287/mksc.2016.1004 .