The authors address two significant challenges in using online text reviews to obtain finegrained attribute level sentiment ratings. First, in contrast to methods that rely on word frequency, they develop a deep learning convolutional-LSTM hybrid model to account for language structure. The convolutional layer accounts for spatial structure (adjacent word groups or phrases) and LSTM accounts for sequential structure of language (sentiment distributed and modified across non-adjacent phrases). Second, they address the problem of missing attributes in text in constructing attribute sentiment scores—as reviewers write only about a subset of attributes and remain silent on others. They develop a model-based imputation strategy using a structural model of heterogeneous rating behavior. Using Yelp restaurant review data, they show superior attribute sentiment scoring accuracy with their model. They find three reviewer segments with different motivations: status seeking, altruism/want voice, and need to vent/praise. Reviewers write to inform and vent/praise, but not based on attribute importance. The heterogeneous model-based imputation performs better than other common imputations; and importantly leads to managerially significant corrections in restaurant attribute ratings. More broadly, our results suggest that social science research should pay more attention to reduce measurement error in variables constructed from text.
We propose an instrumental-variable (IV) approach to estimate the causal effect of service satisfaction on customer loyalty, by exploiting a common source of randomness in the assignment of service employees to customers in service queues. Our approach can be applied at no incremental cost by using routine repeated cross-sectional customer survey data collected by firms. The IV approach addresses multiple sources of biases that pose challenges in estimating the causal effect using cross-sectional data: (i) the upward bias from common-method variance due to the joint measurement of service satisfaction and loyalty intent in surveys; (ii) the attenuation bias caused by measurement errors in service satisfaction; and (iii) the omitted-variable bias that may be in either direction. In contrast to the common concern about the upward common-method bias in the estimates using cross-sectional survey data, we find that ordinary-least-squares (OLS) substantially underestimates the casual effect, suggesting that the downward bias due to measurement errors and/or omitted variables is dominant. The underestimation is even more significant with a behavioral measure of loyalty–where there is no common methods bias. This downward bias leads to significant underestimation of the positive profit impact from improving service satisfaction and can lead to under-investment by firms in service satisfaction. Finally, we find that the causal effect of service satisfaction on loyalty is greater for more difficult types of services.
At many firms, incentivized salespeople with private information about customers are responsible for CRM. While incentives motivate sales performance, private information can induce moral hazard by salespeople to gain compensation at the expense of the firm. We investigate the sales performance–moral hazard tradeoff in response to multidimensional performance (acquisition and maintenance) incentives in the presence of private information. Using unique panel data on customer loan acquisition and repayments linked to salespeople from a microfinance bank, we detect evidence of salesperson private information. Acquisition incentives induce salesperson moral hazard leading to adverse customer selection, but maintenance incentives moderate it as salespeople recognize the negative effects of acquiring low-quality customers on future payoffs. Critically, without the moderating effect of maintenance incentives, adverse selection effect of acquisition incentives overwhelms the sales enhancing effects, clarifying the importance of multidimensional incentives for CRM. Reducing private information (through job transfers) hurts customer maintenance, but has greater impact on productivity by moderating adverse selection at acquisition. The paper also contributes to the recent literature on detecting and disentangling customer adverse selection and customer moral hazard (defaults) with a new identification strategy that exploits the time-varying effects of salesperson incentives.
In many firms, incentivized salespeople with private information about their customers are responsible for customer relationship management (CRM). Private information can help the firm by increasing sales efficiency, but it can also hurt the firm if salespeople use it to maximize own compensation at the expense of the firm. Specifically, we consider two negative outcomes due to private information — ex-ante customer adverse selection at the time of acquisition and ex-post customer moral hazard after acquisition. This paper investigates potential positive and negative responses of a salesforce to managerial levers — multidimensional incentives for acquisition and retention performance and job transfers that affect the level of private information.
Salespeople are responsible for managing customer relationships and compensated through multidimensional performance incentives for customer acquisition and maintenance at many firms. This paper investigates how a salesperson’s private information on customers affect their response to multiple dimensions of incentives. Using unique matched panel data that links individual salesperson performance metrics with customer level loans and repayments from a microfinance bank, we find that sales people indeed possess private information that is not available to the firm. Salespeople use the private information to engage in adverse selection of customers in response to acquisition incentives. Customer maintenance incentives serve a dual purpose; they not only reduce loan defaults, but also moderate adverse selection in customer acquisition. Transfers that eliminate private information reduces the adverse selection effects of acquisition incentives, but increase loan defaults — customer moral hazard. Despite the potential negative adverse selection effects due to private information, the effort increasing effect of each of the three dimensions of sales management we investigate — acquisition incentive, maintenance incentive and transfers all have a net positive effect on firm value. Methodologically, the paper introduces an identification strategy to separate customer adverse selection and customer moral hazard (loan repayment), by leveraging the multidimensional incentives of an intermediary (salesperson) responsible for both customer selection and repayment with private information about customers.
We test for the long-term effects of experience during youth on consumption in nontraditional taste-forming categories. A unique dataset that tracks individuals over twenty years from 1992-2011, residing in nine Chinese provinces that vary widely in both income levels and rate of economic growth, helps us identify cohort and intra-cohort “prosperity-inyouth” (PIY) effects on consumption. We first demonstrate that non-traditional category consumption increases strongly among cohorts that entered adulthood during China’s boom years. We then show evidence of the intra-cohort PIY effect, controlling for individual level experience by leveraging the heterogeneity in the timing and rate of growth in prosperity across Chinese provinces. We find that the PIY effect has two dimensions– a direct effect of one’s own prosperity and an indirect effect of the prosperity of one’s province during youth. The indirect effects suggest that norms and aspirations created by the consumption of nontraditional categories by the surrounding rich during one’s youth have significant impact on long-term consumption—almost the same magnitude as the direct effect. We conduct a large number of robustness checks; in particular, we rule out potential supply side and attitude based explanations for the PIY effect. Our results imply that segmentation and consumption forecasts based on birth cohorts and experience of prosperity can be effective for taste forming non-traditional products in emerging markets.
We test for the long-term impact of experiencing “prosperity in youth” (PIY) on non-traditional category consumption. Using unique twenty-year panel data of individuals from nine Chinese provinces with varying levels of per-capita GDP and rates of per-capita GDP growth, we find robust evidence for the PIY effect. We find both a direct effect of one’s own prosperity and an indirect effect of the prosperity of one’s province during youth on long-term consumption. In particular, the indirect PIY effect is driven more strongly by individuals with low incomes during youth — suggesting that norms and aspirations created by the consumption of non-traditional categories by the rich during one’s youth have significant impact on long-term consumption — almost the same magnitude as the direct effect. The analysis also highlights the importance of separating cohort effects from life cycle effects for taste based products. We highlight the marketing implications for non-traditional categories in emerging markets.
Firms make investments in technology to increase productivity. But in emerging markets, where a culture of informality is widespread, information technology (IT) investments leading to greater transparency can impose a cost through higher taxes and need for regulatory compliance. This tendency of firms to avoid productivity-enhancing technologies and remain small to avoid transparency has been dubbed the “Peter Pan Syndrome.” We examine whether firms make the tradeoff between productivity and transparency by examining IT adoption in the Indian retail sector. We find that computer technology adoption is lower when firms have motivations to avoid transparency. Specifically, technology adoption is lower when there is greater corruption, but higher when there is better enforcement and auditing. So firms have a higher productivity gain threshold to adopt computers in corrupt business environments with patchy and variable enforcement of the tax laws. Not accounting for this motivation to hide from the formal sector underestimates productivity gains from computer adoption. Thus in addition to their direct effects on the economy, enforcement, auditing and corruption can have indirect effects through their negative impact on adoption of productivity enhancing technologies that also increase operational transparency.
In retail settings with price promotions, consumers often search across stores and time. However the search literature typically only models one pass search across stores, ignoring revisits to stores; the choice literature using scanner data has modeled search across time, but not search across stores in the same model. We develop a multi-pass search model that jointly endogenizes search in both dimensions; our model nests a nite horizon model of search across stores within an in nite horizon model of inter-temporal search. We apply our model to milk purchases at grocery stores; hence the model also accounts for repeat purchases across time, inventory holding by households and grocery basket effects. We note that the special case without these additional features can be used to study one time purchases with repeat store visits as in the case of durable goods and online shopping. We formulate the empirical model as a mathematical program with equilibrium constraints (MPEC) and estimate it allowing for latent class heterogeneity using an iterative E-M algorithm. In contrast to extant research, we nd that omitting the temporal dimension underestimates price elasticity. We attribute this difference to the relative frequency of household stockouts and purchase frequency in the milk category. Interestingly, increasing the promotional frequency (while reducing its depth to maintain the mean and variance of prices across all stores) can increase loyalty to the household’s preferred store.
We randomize advertising content motivated by the psychology literature on sympathy generation and framing effects in mailings to about 185,000 prospective new donors in India. We find significant impact on the number of donors and amounts donated consistent with sympathy biases such as the “identifiable victim,” “in-group” and “reference dependence.” A monthly reframing of the ask amount increases donors and amount donated relative to daily reframing. A second field experiment targeted to past donors, finds that the effect of sympathy bias on giving is smaller in percentage terms but statistically and economically highly significant in terms of the magnitude of additional dollars raised. Methodologically, the paper complements the work of behavioral scholars by adopting an empirical researchers’ lens of measuring relative effect sizes and economic relevance of multiple behavioral theoretical constructs in the sympathy bias and charity domain within one field setting. Beyond the benefit of conceptual replications, the effect sizes provide guidance to managers on which behavioral theories are most managerially and economically relevant when developing advertising content.
In response to price dispersion across stores and price promotions over time, consumers search across both stores (spatial) and time (temporal), in many retail settings. Yet there is no search model in extant research that jointly endogenizes search in both dimensions. We develop a model of spatiotemporal search that nests a finite horizon model of spatial search across stores within an infinite horizon model of inter-temporal search. The model is estimated using an iterative procedure that formulates it as a mathematical program with equilibrium constraints (MPEC) embedded within an E-M algorithm to allow estimation of latent class heterogeneity. The empirical analysis uses data on household store visits and purchases in the milk category. In contrast to extant research, we find that omitting the temporal dimension underestimates price elasticity. We attribute this difference to the relative frequency of household stock outs and purchase frequency in the milk category. Further, contrary to the conventional wisdom that promotions increase store switching and reduces store loyalty, we find that in the presence of search frictions, price promotions can be a store loyalty-enhancing tool.
In response to price dispersion across stores and price promotions over time, consumers search across both stores and time, in many retail settings. Yet there is no search model in extant research that jointly endogenizes search in both dimensions. We develop a model of search across stores and across time that nests a nite horizon model of search across stores within an in nite horizon model of inter-temporal search. The model is estimated using an iterative procedure that formulates it as a mathematical program with equilibrium constraints (MPEC) embedded within an E-M algorithm to allow estimation of latent class heterogeneity. The empirical analysis uses data on household store visits and purchases in the milk category. In contrast to extant research, we nd that omitting the temporal dimension underestimates price elasticity. We attribute this difference to the relative frequency of household stockouts and purchase frequency in the milk category. Interestingly, we nd that for a given mean and variance in price, increasing the frequency while reducing depth of price promotions across all stores can increase share of visits and pro ts for consumers’ preferred store for the segment with high price sensitivity and low cross-store search cost; consumers who are most prone to search across stores and across time.
This paper offers a new identification strategy for disentangling structural state dependence from unobserved heterogeneity in preferences. Our strategy exploits market environments where there is a choice-consumption mismatch. We first demonstrate the effectiveness of our identification strategy in obtaining unbiased state dependence estimates via Monte Carlo analysis and highlight its superiority relative to the extant choice-set variation based approach. In an empirical application that uses data of repeat transactions from the car rental industry, we find evidence of structural state dependence, but show that state dependence effects may be overstated without exploiting the choice-consumption mismatches that materialize through free upgrades.
This paper investigates the impact of spatial zoning restrictions on retail market outcomes. We estimate a structural model of entry, location and format choice across a large number of markets in the presence of zoning restrictions. The paper contributes to the literature in three ways: First, the paper demonstrates that the omission of zoning restrictions in the extant literature on entry and location choice leads to biased estimates of the factors affecting market potential and competitive intensity. Second, the cross-market variations in zoning regulations helps us test and provide evidence for the theory that constraints on spatial differentiation will lead to greater product differentiation. Finally, we provide qualitative insight on how zoning impacts retail entry and format variety; in particular we evaluate the impact of prototypical zoning arrangements such as “centralized,” “neighborhood,” and “outskirt” zoning on entry and format variety.
Beverage consumption occurs many times a day in response to a variety of needs that change throughout the day. In making their choices, consumers self-regulate their consumption by managing short run needs (e.g., hydration and mood pickup) with long-term goals (e.g., health). Using unique intra-day beverage consumption, activity and psychological needs data, we develop and estimate a model of high frequency consumption choices that accounts for both intra-day changes in short run needs and individual level unobserved heterogeneity in the degree of self-regulation. A novel feature of the model is that it allows for dynamics of consumption and stockpiling at the level of product attributes. The model is used to evaluate introduction of new products in the beverage category and gain insight into the linkage between self-regulation and excess consumption. Broadly, the modeling framework of balancing short run needs with long-term goals has wide ranging applications in choices where long term effects are gradual (e.g., nutrition, exercise, smoking and preventive health care).
We illustrate an approach to measure demand externalities from co-location by estimating household level changes in grocery spending at a supermarket among households that also buy gas at a co-located gas station, relative to those who do not. Controlling for observable and unobserved selection in the use of gas station, we find significant demand externalities; on average a household that buys gas has 7.7% to 9.3% increase in spending on groceries. Accounting for differences in gross margins, the profit from the grocery spillovers is 130% to 150% the profit from gasoline sales. The spillovers are moderated by store loyalty, with the gas station serving to cement the loyalty of store-loyal households. The grocery spillover effects are significant for traditional grocery products, but 23% larger for convenience stores. Thus co-location of a new category impacts both inter-format competition with respect to convenience stores (selling the new category) and intra-format competition with respect to other supermarkets (selling the existing categories).