Sam K. Hui

The Effect of In-Store Travel Distance on Unplanned Spending: Applications to Mobile Promotion Strategies

Typically, shoppers’ paths only cover less than half of the areas in a grocery store. Given that shoppers often use physical products in the store as external memory cues, encouraging shoppers to travel more of the store may increase unplanned spending. Estimating the direct effect of in-store travel distance on unplanned spending, however, is complicated by the difficulty of collecting in-store path data and the endogeneity of in-store travel distance. To address both issues, the authors collect a novel data set using in-store radio frequency identification tracking and develop an instrumental variable approach to account for endogeneity. Their analysis reveals that the elasticity of unplanned spending on travel distance is 57% higher than the uncorrected ordinary least squares estimate. Simulations based on the authors’ estimates suggest that strategically promoting three product categories through mobile promotion could increase unplanned spending by 16.1%, compared with the estimated effect of a benchmark strategy based on relocating three destination categories (7.2%). Furthermore, the authors conduct a field experiment to assess the effectiveness of mobile promotions and find that a coupon that required shoppers to travel farther from their planned path resulted in a substantial increase in unplanned spending (

Research Note---The Traveling Salesman Goes Shopping: The Systematic Deviations of Grocery Paths from TSP Optimality

21.29) over a coupon for an unplanned category near their planned path (

Spatiotemporal analysis of imitation behavior across new buyers at an online grocery retailer

13.83). The results suggest that targeted mobile promotions aimed at increasing in-store path length can increase unplanned spending.

Deconstructing the “First Moment of Truth”: Understanding Unplanned Consideration and Purchase Conversion Using In-Store Video Tracking

We examine grocery shopping paths using the traveling salesman problem (TSP) as a normative frame of reference. We define the TSP-path for each shopper as the shortest path that connects all of his purchases. We then decompose the length of each observed path into three components: the length of the TSP-path, the additional distance because of order deviation (i.e., not following the TSP-order of category purchases), and the additional distance because of travel deviation (i.e., not following the shortest point-to-point route). We explore the relationship between these deviations and different aspects of in-store shopping/purchase behavior. Among other things, our results suggest that (1) a large proportion of trip length is because of travel deviation; (2) paths that deviate substantially from the TSP solution are associated with larger shopping baskets; (3) order deviation is strongly associated with purchase behavior, while travel deviation is not; and (4) shoppers with paths closer to the TSP solution tend to buy more from frequently purchased product categories.

Modeling DVD Preorder and Sales: An Optimal Stopping Approach

For Internet retailers, demand propagation varies not only through time but also over space. The authors develop a Bayesian spatiotemporal model to study two imitation effects in the evolution of demand at an Internet retailer. Building on previous literature, the authors allow imitation behavior to be reflected both in geographic proximity and in demographic similarity. As these imitation effects can be time varying, the authors specify their dynamics using a “polynomial smoother” embedded within the Bayesian framework. They apply the model to new buyers at Netgrocer. com and calibrate it on 45 months of data that span all 1459 zip codes in Pennsylvania. The authors find that the proximity effect is especially strong in the early phases of demand evolution, whereas the similarity effect becomes more important with time. Over time, new buyers are increasingly likely to emerge from new zip codes beyond the “core set” of zip codes that produce the early new buyers, and spatial concentration declines. The authors explore the managerial implications stemming from these findings through a hypothetical “seeding” experiment. They also discuss other implications for Internet retailing practice.

Assessing Box Office Performance Using Movie Scripts: A Kernel-Based Approach

Retailers and manufacturers are keenly interested in understanding unplanned consideration and purchase conversion, but data that capture in-store product consideration have been unavailable in the past. In the current research, the authors use in-store video tracking to collect a novel data set that records shopping behavior at the point of purchase, including product consideration. In conjunction with an entrance survey of purchase intentions, they conduct several descriptive analyses that focus on the incidence, category propensity, behavioral characteristics, and outcome of unplanned consideration. The results reveal several new empirical insights. First, the authors find significant category-level complementarities between planned items and unplanned considerations, which they capture using a latent category map. Second, planned consideration and unplanned consideration differ in key behavioral characteristics (e.g., likelihood of purchase, time of occurrence, number of product touches). Third, greater likelihood of purchase conversion is significantly associated with dynamic factors (e.g., remaining in-store slack, outcome of the previous consideration) and behavioral characteristics (e.g., number of displays viewed, distance to shelf, references to a shopping list). The authors conclude with a discussion of implications of these findings for research and shopper marketing.

Decision Models for the Movie Industry

When a DVD title is announced prior to actual distribution, consumers can often preorder the title and receive it as soon as it is released. Alternatively, once a title becomes available i.e., formally released, consumers can obtain it upon purchase with minimal delay. We propose an individual-level behavioral model that captures the aggregate preorder/postrelease sales of motion picture DVDs. Our model is based on an optimal stopping framework. Starting with the utility function of a forward-looking consumer, and allowing for consumer heterogeneity, we derive the aggregate preorder/postrelease sales distribution. Even under a parsimonious specification for the heterogeneity distribution, our model recovers the typically observed temporal pattern of DVD preorder and sales, a pattern which exhibits an exponentially increasing number of preorder units before the release, peaks at release, and drops exponentially afterward. Using data provided by a major Internet DVD retailer, we demonstrate a number of important managerial implications stemming from our model. We investigate the role of preorder timing through a policy experiment, estimate residual sales, and forecast post-release sales based only on preorder information. We show that our model has substantially better predictive validity than benchmark models.

The Role of Surprise: Understanding Overreaction and Underreaction to Unanticipated Events using In-Play Soccer Betting Market

We develop a methodology to predict box office performance of a movie at the point of green-lighting, when only its script and estimated production budget are available. We extract three levels of textual features (genre and content, semantics, and bag-of-words) from scripts using screenwriting domain knowledge, human input, and natural language processing techniques. These textual variables define a distance metric across scripts, which is then used as an input for a kernel-based approach to assess box office performance. We show that our proposed methodology predicts box office revenues more accurately (29 percent lower mean squared error (MSE)) compared to benchmark methods.

Bayesian Multi-Resolution Spatial Analysis with Applications to Marketing

The movie industry represents a challenging domain for scholarly research in general and for modelers in particular. The industry is characterized by a product (content) with multiple distribution outlets, each having a relatively short window of opportunity (see Fig. 13.1). Distribution outlets for movies include domestic and foreign theatres, home video, cable TV, and network TV. In each of these windows, many different parties and decision makers are involved.While some emphasize the creative and artistic aspects of the product, others focus on the business issues. Some industry pundits claim that the artistic nature of the product and its uncertain quality make the movie industry inherently different from others; hence, any formal methods or models employed in other industries to improve operational and commercial performance are irrelevant. Furthermore, unlike other industries where trends and consumer tastes are tracked continuously, studios see themselves more as trend setters than as trend followers. Views of industry experts are divided when it comes to reasons for recent box office declines: some blame it on deteriorating product qualities, others on changes in consumer behavior. Our experiences and perspectives are different from those who argue that movie is a form of art that cannot be understood with formal quantitative methods. Rather, we think that there is a creative tension between art and science that, if balanced properly, can lead to improvement in both the business and the artistic spheres of the industry. Decision-making style varies across the different parties involved in the production and distribution of movies. Film makers, coming from artistic backgrounds, tend to believe in more intuitive styles. In contrast, executives in the home video sector, who interact more closely with retailers and consumers, generally see more value in formal decision models. The rise of a new breed of business-educated executives, who are starting to fill high

Assessing the Impact of Peer Educator Outreach on the Likelihood and Acceleration of Clinic Utilization among Sex Workers.

Previous research in finance has found evidences of both overreaction and underreaction to unanticipated events, but has yet to explain why investors overreact to certain events while underreacting to others. In this paper, we hypothesize that while market participants generally underreact to new events due to conservatism, the extent of underreaction is moderated by “surprise,” thus causing market participants to overreact to events that are highly surprising. We test our hypothesis using data from an in-play soccer betting market, where new events (goals) are clearly and exogenously defined, and the degree of “surprise” can be directly quantified (goals scored by underdogs are more surprising). We provide both statistical and economic evidences in support of our hypothesis.We test the relationship between over-(under)reaction to news and the degree of surprise. Investors/bettors generally rely too heavily on prior beliefs and underreact, but sufficiently surprising information may cause overreaction due to the salience of the news. We use data from in-play soccer betting, where the arrival of information (goals) is well defined and the degree of surprise can be quantified. We find that bettors underreact to most goals, but overreact to highly surprising goals scored by “underdogs.” Our analysis suggests that overand underreactions are subsequently corrected, and a profitable betting strategy could be formed based on the biases.