Andres Musalem

Structural Estimation of the Effect of Out-of-Stocks

We develop a structural demand model that endogenously captures the effect of out-of-stocks on customer choice by simulating a time-varying set of available alternatives. Our estimation method uses store-level data on sales and partial information on product availability. Our model allows for flexible substitution patterns, which are based on utility maximization principles and can accommodate categorical and continuous product characteristics. The methodology can be applied to data from multiple markets and in categories with a relatively large number of alternatives, slow-moving products, and frequent out-of-stocks (unlike many existing approaches). In addition, we illustrate how the model can be used to assist the decisions of a store manager in two ways. First, we show how to quantify the lost sales induced by out-of-stock products. Second, we provide insights on the financial consequences of out-of-stocks and suggest price promotion policies that can be used to help mitigate their negative economic impact, which run counter to simple commonly used heuristics.

Who's Got the Coupon? Estimating Consumer Preferences and Coupon Usage from Aggregate Information

Most researchers in marketing have typically relied on disaggregate data (e.g., consumer panels) to estimate the behavioral and managerial implications of coupon promotions. In this article, the authors propose the use of individual-level Bayesian methods for studying this problem when only aggregate data on consumer choices (market share) and coupon usage (number of distributed coupons and/or number of redeemed coupons) are available. The methodology is based on augmenting the aggregate data with unobserved (simulated) sequences of choices and coupon usage consistent with the aggregate data. The authors analyze various marketing scenarios that differ in terms of their assumptions about consumer choices, coupon availability, and coupon redemption. They illustrate the proposed methods using both simulated data and a real data set for which an extensive set of posterior predictive checks helps validate the aggregate-level estimation. In addition, the authors relate the empirical results to some of the findings in the literature about the coordination of coupon promotions and pricing and show how the methodology can be used to evaluate alternative coupon targeting policies.

Eye-Tracking Reveals Processes that Enable Conjoint Choices to Become Increasingly Efficient with Practice

Choice-based conjoint is a popular technique for characterizing consumers’ choices. Three eye-tracking studies explore decision processes in conjoint choices that take less time and become more accurate with practice. These studies reveal two simplification processes that are associated with greater speed and reliability. Alternative focus gradually shifts attention toward options that represent promising choices, whereas attribute focus directs attention to important attributes that are most likely to alter or confirm a decision. Alternative and attribute focus increase in intensity with practice. In terms of biases, the authors detect a small but consistent focus on positive aspects of the item chosen and negative aspects of the items not chosen. They also show that incidental exposures arising from the first-examined alternative or from alternatives in a central horizontal location increase attention but have a much more modest and often insignificant impact on conjoint choices. Overall, conjoint choice is found to be a process that is (1) largely formed by goal-driven values that respondents bring to the task and (2) relatively free of distorting effects from task layout or random exposures.

Underpromising and Overdelivering: Strategic Implications of Word of Mouth

We analyze a firms optimal communication strategy for setting consumer expectations when consumers are uncertain about product quality and word of mouth is prevalent in the market. We derive three main results: [i] Extant signaling theory argues that advertising should be costly for it to be informative of product quality. We show that in the presence of negative word of mouth, even costless advertising can serve as an informative signal for quality. [ii] Conventional wisdom suggests that as the consequences of negative word of mouth become stronger, a firm should become more cautious in setting high consumer expectations, to prevent future disappointment. We show that this need not always be the case: interestingly, when negative word of mouth is prevalent and its consequences become stronger, a firm might become more aggressive in setting high expectations given consumer rationality. [iii] Disconfirmation (defined as a stronger shift in beliefs when experiences are inconsistent with messages) serves as an adequate mechanism for preventing quality misrepresentation by a firm in its communications to consumers. But when markets are characterized by confirmation effects (a tendency to discount experiences inconsistent with messages) and future sales are important, we observe firms might entirely misrepresent their quality.

Retail in High Definition: Monitoring Customer Assistance through Video Analytics

Abstract Staffing decisions typically account for a large portion of a retailers operational costs. The effectiveness of these decisions has often been analyzed by relating staffing levels to revenues. However, such approach does not explicitly consider the mechanisms by which the staff can contribute to generate revenues, such as customer assistance. This motivates the development of a fast, efficient, high-frequency method to measure customer assistance in real time. The method relies on the use of short videos that track only a portion of a customers shopping path. The recorded videos may not track all the relevant information to identify a customer-employee interaction, i.e. they might be censored. Accordingly, we develop a survival model to analyze these data, defining unbiased estimates of customer assistance. This methodology also gives insights into how staffing decisions translate into different levels of customer assistance under different congestion scenarios. For example, when the store is congested, increasing the staff from one to four employees can increase the fraction of customers receiving assistance from 38% to 45%. Furthermore, these assistance rate measures can in turn be used to assess the economic impact of assisting customers in terms of conversion or basket size. This introduces important estimation challenges related to the endogeneity of customer assistance (e.g., if the customers that are more likely to purchase are also more likely to seek assistance) and the measurement error in customer assistance rates. We address both issues using an instrumental variables approach that relies on variations on service capacity as a driver of exogenous variance in customer assistance. In particular, we find that raising the assistance rate from 50% to 60% (a one standard deviation increase from the average) increases conversion by about 5 percentage points, corresponding to a 18.5% increase in transaction volume. Finally, we show that the approach developed in this work is useful to support store staffing decisions.

When Demand Projections are Too Optimistic: A Structural Model of Product Line and Pricing Decisions

A methodology is proposed to estimate structural models of product line competition. This methodology enables researchers to estimate demand systems accounting for the endogeneity of the mix of products available in each market, an issue which is typically ignored in the empirical literature. In particular, it is observed that not accounting for this form of endogeneity leads to overoptimistic estimates of total demand due to a sample selection bias. More importantly, biased estimates of demand can generate inaccurate inferences about consumer welfare or imply misleading managerial recommendations.The proposed model jointly considers the interplay between consumer preferences, pricing and assortment decisions. Consumer demand is characterized by a utility maximization process with unobserved heterogeneity in consumer preferences. Price decisions are assumed to be the outcome of a Bertrand-Nash game among firms offering differentiated products. Product line decisions are modeled using a Bayesian-Nash equilibrium concept where firms form beliefs about the profits of their competitors and anticipate the prices and demand they would observe for any given set of products that could be introduced in the market. The estimation approach is implemented relying on parallelization decomposing some of the most computationally intensive steps into a series of independent and much smaller problems.The methodology is illustrated using both simulated and real data, where the latter refers to purchases in the liquid laundry detergent product category. The results show that ignoring this form of endogeneity leads a researcher to overestimate the demand, prices and profits for products that have not yet been introduced in the market, while for existing products market shares and profits are substantially underestimated when performing policy simulations involving the addition of products to current assortments.

A Parsimonious Structural Model of Individual Demand for Multiple Related Goods

Nearly all theoretically motivated models of consumer demand for multiple goods assume additive separability in preferences, i.e. the consumption utility of each good x is independent of the quantity demanded of another good y. This is a strong restriction that makes the solution of the consumer’s utility maximization problem computationally tractable. This paper shows that assuming preferences are weakly separable yields a similar simplification. It offers a theoretically founded model of consumer demand for continuous quantities of related goods. It also proposes a Bayesian estimation approach with a parsimonious parameterization that allows for corner solutions. It is the first structural model of individual demand for multiple goods which relaxes additive separability and does not suffer from a curse of dimensionality in the number of chosen goods. The model is estimated using data on Chilean advertisers’ television audience purchases. We find that advertisers prefer spreading expenditures across time blocks more so than spreading expenditures across networks within a time block. The model nests additively separable preferences as a special case, but the data reject this case. We illustrate how a television network could use the model to assess the consequences of different advertisement pricing policies conditional on competitive response assumptions.

Market basket analysis insights to support category management

This paper aims to present an approach to detect interrelations among product categories, which are then used to produce a partition of a retailer’s business into subsets of categories. The methodology also yields a segmentation of shopping trips based on the composition of each shopping basket.,This work uses scanner data to uncover product category interdependencies. As the number of possible relationships among them can be very large, the authors introduce an approach that generates an intuitive graphical representation of these interrelationships by using data analysis techniques available in standard statistical packages, such as multidimensional scaling and clustering.,The methodology was validated using data from a supermarket store. The analysis for that particular store revealed four groups of products categories that are often jointly purchased. The study of each of these groups allowed us to conceive the retail store under study as a small set of sub-businesses. These conclusions reinforce the strategic need for proactive coordination of marketing activities across interrelated product categories.,The approach is sufficiently general to be applied beyond the supermarket industry. However, the empirical findings are specific to the store under analysis. In addition, the proposed methodology identifies cross-category interrelations, but not their underlying sources (e.g. marketing or non-marketing interrelations).,The results suggest that retailers could potentially benefit if they transition from the traditional category management approach where retailers manage product categories in isolation into a customer management approach where retailers identify, acknowledge and leverage interrelations among product categories.,The authors present a fast and wide-range approach to study the shopping behavior of customers, detect cross-category interrelations and segment the retailer’s business and customers based on information about their shopping baskets. Compared to existing approaches, its simplicity should facilitate its implementation by practitioners.

Does Word of Mouth Reduce Advertising That Signals Quality

We analyze a firm’s optimal communication strategy when dissipative advertising can be used as a signal of unobserved quality for an experience good, consumers share experiences via word of mouth, and word of mouth can be biased. We study the impact of two distinct empirically documented behavioral biases in word of mouth: negativity and positivity. In terms of the first of these biases, a priori, one might expect that with more negative opinions being shared, it should be easier for a low quality firm to be exposed and hence a high quality firm may need a smaller investment to separate itself in the eyes of rational consumers. Surprisingly, we show that with more negativity bias, a high quality firm becomes more aggressive in signaling its quality. This is because when negative word of mouth is prevalent and consumers hear about a negative experience, they are more likely to be forgiving while updating their quality beliefs. This yields important benefits to a low quality firm, and as a consequence, to effectively achieve separation and prevent the low type from mimicking, a high quality firm needs to increase its advertising spending. Such firm behavior crucially relies on followers being aware of the existence and magnitude of this bias; and is reversed otherwise. Similar results hold for positivity bias, when biases arise due to under-reporting, and when a firm can rely on prices to signal quality to consumers along with advertising. Overall, our analysis suggests that as bias in word of mouth increases, it is optimal for a high quality firm to shift to a more aggressive communication strategy.We analyze a firm’s optimal communication strategy when dissipative advertising serves as a signal of quality for an experience good, and consumers share experiences via word of mouth which can be positively or negatively biased. With negativity bias, a priori one might expect that when more negative opinions are shared, a high quality firm should spend less on advertising, since a low type would gain less from mimicking the high type. Surprisingly, we show that a higher advertising spending is needed instead. This is because when negative word of mouth is prevalent and consumers hear about a negative experience, they are forgiving while updating their quality beliefs. This behavior benefits the low quality firm; hence to prevent a low type from mimicking, a high type increases its advertising spending. Analogous results hold for positivity bias, and when a firm uses prices to signal its quality along with advertising.