Current Research

Click on any topic below to read more information.

1. An Empirical Study of Healthy Consumption.

Research Team: Cheng-Ting Chen (Undergraduate student, University at Buffalo), Vijay Ganesh Hariharan (PhD student, University at Buffalo) and Minakshi Trivedi (University at Buffalo).

Details: Part of marketing research is to understand the consumption patterns in the consumer market, with regards to healthy and unhealthy products. The information to be derived from such research will be valuable to the producers of food and beverages that may have certain health risks associated with it. In this study, using the Buffalo CRM data base, we empirically look at differences in patterns of consumption for several product categories across geographic regions and demographics.

2. A Field Examination of the Influence of Brand Equity on Behavioral Loyalty and Factors that Moderate this Relationship.

Research Team: Kalpesh Desai (University at Binghamton), Jan Hofmeyr, Dr. J. Jeffrey Inman (University of Pittsburgh) and Debabrata Talukdar (University at Buffalo). Authors listed alphabetically.

Details: While prior research has highlighted the critical role of brand equity in influencing customer retention, the mapping between brand equity and customer retention has been less than perfect. In this research, we posit that this linkage is moderated by individual difference variables of brand equity of other alternatives in the category, importance of brand choice, and ambivalence (arising from perceived low differentiation between alternatives). Across several product categories, we will relate consumer’s purchases (available from scanner data) with their attitudinal data (to be collected through surveys) about the equity of distinct brands in the category. We will measure at individual level nine dimensions of brand equity and the individual difference variables specified above along with other factors that moderate the influence of brand equity on brand choice. This will enable us to test hypotheses about consumer moderators responsible for heterogeneity in the brand equity – choice linkage. In addition, it will facilitate testing hypotheses about the contexts in which the influence of brand equity on choice is higher and also investigate the specific dimension(s) of brand equity that result in the enhanced influence of brand equity on choice in those contexts. We define these contexts with respect to brand factors (e.g., the market leader, the private label brand), household factors (e.g., heavy category users, older consumers), and marketing mix factors (e.g., price promotion of non-target brands).

3. Implications of Consumer Variety Seeking for Retailers’ Product Line Decisions.

Research Team: Kalpesh Desai (University at Binghamton), P. B. (Seethu) Seetharaman (Rice University) and Minakshi Trivedi (University at Buffalo). Authors listed alphabetically.

Details: We investigate the consequences of consumers’ variety-seeking along product attributes for retailers’ product line decisions. Using scanner and survey data from a sample of households in the database of a local grocery chain, we estimate a demand model that is based on the assumption that consumers switch between brands over time in response to their desire for variety in the attributes being consumed. Using the estimates of the demand model, we then perform managerial policy simulations that show the degree to which retail profits can be improved by appropriately extending or pruning the length of product lines in response to the documented variety-seeking effects in demand. In doing so, we are able to explicitly disentangle the separate roles of (i) across-consumer heterogeneity in preferences for product attributes (which are constant over time), and (ii) within-consumer variety-seeking in product attributes over time. We plan to confirm some of the findings from the above model estimation by running lab experiments.

4. Context-Specific Positive Influence of Walmart on Value-Priced Grocery Stores.

Research Team: Kalpesh Desai (University at Binghamton), Karthik Sridhar (PhD student, University at Buffalo), Andrei Strijnev (University of Texas, Dallas) and Debabrata Talukdar (University at Buffalo). Authors listed alphabetically.

Details: A widespread expectation of Walmart’s entry into grocery business is that it will have a uniformly adverse impact on the businesses of grocery stores, both across products types and types of grocery stores. Using the literature on context effects as theoretical underpinning, we propose that Walmart’s entry may produce beneficial effects for value-priced grocery stores in those markets where prior to the entry of Walmart, these stores were primarily competing against a more upscale grocery store. The positive influence will be restricted to products with high price-high quality associations (e.g., wine, flowers, expensive bakery products) and private label products (including produce) because the low price-low quality associations for these products that were earlier linked to value-priced grocery stores will now shift to Walmart.

5. Does Ingredient Branding Help Choice of Host and Ingredient Brands?

Research Team: Kalpesh Desai (University at Binghamton), Dinesh Kumar Gauri (PhD student, University at Buffalo), Yu Ma (University of Alberta) and S. Ratneshwar (University of Missouri, Columbia).

Details: In this study we investigate if and how both host and ingredient brands benefit from the ingredient branding alliance. After controlling for some covariates, our model specifically ascertains if the sale of ingredient brand, post incorporation into the host brand, increased, decreased, or remained steady. Moreover, if it improved, what is the source – current ingredient brand consumers using the ingredient brand more and/or new consumers (how many are host brand users vs. non host brand users)? On the other hand, if the ingredient brand sales declined, is it because current consumers of ingredient brand are now “satisfying” their need for the ingredient product through the use of host brand? Our model ascertains answers to similar questions for the host brand and investigates the role of three moderating variables.

6. Benchmarking Performance in the Retail World: An Integrated Approach.

Research Team: Dinesh Kumar Gauri (PhD student, University at Buffalo), Gabor Pauler (University of Pécs, Hungary) and Minakshi Trivedi (University at Buffalo).

Details: Standardizing performance expectations across different stores within a chain, differing in store features, its consumers, and the nature of competition it faces, can be an onerous task. We develop an integrated, non-linear, market share model of store expectations that draws upon the existing trade area as well as store performance literatures. By incorporating and normalizing a large number of external and internal factors impacting performance, we are able to offer a means for the retailer to determine equitable standards. The model is estimated using Maximum Likelihood Estimation, on a data set fashioned from several sources. Finally, we propose a set of indices that allows us to evaluate relative performances of stores and regions given the competitive environments they face.

7. Variety Seeking and the Two Stage Model of Choice.

Research Team: Minakshi Trivedi (University at Buffalo), Amresh Kumar (PhD student, University at Buffalo) and Kalpesh Desai (University at Binghamton).

Details: In this study we investigate the notion that since high variety seekers prefer and seek variety, they are likely to perceive differences between two options that low variety seekers might not notice. Combining scanner and survey data, we examine this issue in the context of the two stage model of choice — brand consideration and brand evaluation resulting in choice (Nedungadi 1990). When translated into the two stage choice model, the above seeking of differences suggests that if high variety seekers seek variety at the consideration stage, they will consider very different options that are still perceived to satisfy their goals i.e., they will exaggerate similarities among the very different options compared to low variety seekers. In contrast, if they seek variety at choice stage, they will exaggerate differences among the chosen options compared to low variety seekers.

8. Optimizing Purchase Behavior at the Basket Level.

Research Team: Vijay Ganesh Hariharan (PhD student, University at Buffalo) and Minakshi Trivedi (University at Buffalo).

Details: Multi-category choice models are an increasingly popular class of models, whereby consumer preferences for brands in multiple categories are modeled using a joint distribution that allows different categories to be correlated. Thus, for example, when choosing a salty snack, researchers will typically analyze, say, potato chips, pretzels and nachos. We argue, however, that in many cases the study of consumption behavior cannot be constrained to a limited number of categories. We plan to show, using a multi-category, latent structure model, corroborated by survey data to explore the ‘whys’, that consumers optimize utility not just over a few similar categories, but in fact, over a strategically pre-selected set of categories, and the entire shopping basket. Such basket level utility maximizing behavior has tremendous implications for retail and manufacturer strategy.

9. Disaggregating the Promotional Bump.

Research Team: Karthik Sridhar (PhD student, University at Buffalo), Sanjib Mohanty (PhD student, University of Rochester) and Minakshi Trivedi (University at Buffalo).

Details: In this paper, we disaggregate the promotional impact to model separately the advantages to the retailer and the manufacturer. Using an integrated model for purchase behavior, we evaluate the impact of promotions on profitability measures and address issues such as financial implications for category expansion or purchase feedback effect.

10. Scaling Response Parameters: A Multivariate Probit Approach.

Research Team: Sri Devi Duvvuri (University of Iowa) and Minakshi Trivedi (University at Buffalo).

Details: The goal of this research is to compare consumer response parameters across categories and regions to study implications of different variances in mean responses while specifying random coefficient choice models. In recent years, there has been an interest in exploring this assumption (Train 2003, Louviere et. al. 2005) with Swait and Louviere (1993) pointing out that a comparison of estimated parameters across segments and/or data sets might confound the results. In this research, we explore these issues using a multivariate probit model to estimate response sensitivities. We use scanner panel data across several categories from different regions to study these effects.

11. Consumer Preference Evolution in New Product Categories: A Multivariate Analysis of Organic Brands.

Research Team: Ram Bezawada (University at Buffalo), Minakshi Trivedi (University at Buffalo) and Deanna Wang (San Francisco State University).

Details: In this paper, we seek to understand how consumer perceptions and attitudes towards new products evolve using both survey data as well as actual purchases through scanner panel data. We thus relate attitudinal data and behavioral intent data with revealed preference data, and assess the extent of consumer preference evolution in new products at both category level and brand level. We use the organic food market which has grown rapidly over the over the past few years and was estimated to be around $20 billion in 2005, to test our model.

12. An Empirical Analysis of the Determinants of the Pricing Strategy and Format of a Retail Store.

Research Team: Dinesh Kumar Gauri (PhD student, University at Buffalo) and Minakshi Trivedi (University at Buffalo).

Details: Two powerful and highly effective strategic tools that retailers possess relate to pricing and store format decisions. From the several choices available for each decision, a retailer can choose any combination. Past research has focused mainly on consumer response to such pricing and/or formatting strategies. Surprisingly, particularly given the nature of the competitive environment the retailer faces, not much attention has been given to the issue of what makes the retailer choose a particular price/format combination in the first place. This then, is the focus of our work. Given that any price/format structure would be specific to a location and face a specific set of competitive conditions, regional and competitive factors would obviously play an important part in determining a particular price/format strategy. Based on our unique data set covering all the grocery retailers across six states, we use a multinomial logit model to study the determinants of the price/format strategy for the retailers. Results show that some combinations are more similar than others, and that the competitor’s strategy strongly influences the strategy for the entrant.

13. The Impact of Multi-Category Purchasing: The More the Merrier?

Research Team: Minakshi Trivedi (University at Buffalo)

Details: Multi category shopping by consumers is a well researched phenomenon, particularly in recent years when a combination of greater data availability and increased computing power has made the analysis far more interesting as well as much less time consuming. Not much of this research, however, looks into the issue of multi category purchasing from the retailer’s point of view, nor into the inherent presumption of ‘the more categories, the better’. Particularly when cross category purchasing is frequently motivated through promotions and rewards, additional category sales are often made at less than ‘full price.’ Is the cost of inducing consumers into additional category purchasing, in a situation where the average consumer carries 7 store (loyalty) cards and participates in several frequency programs simultaneously, too high? Is there then, a certain number or range of categories that retailers would find it optimal to reach for, beyond which the benefits decline and the opportunity cost of the promotional investment becomes too high? We propose a dynamic optimization model with the retailer’s objective being to maximize customer profitability over various categories.

14. Cumulative Sum Methods for Spatial Surveillance.

Research Team: Peter A. Rogerson (University at Buffalo), Minakshi Trivedi (University at Buffalo) and Sharmistha Bagchi-Sen (University at Buffalo)

Details: There is often interest in monitoring health within a study region where data are available for a number of sub regions. One way to carry out monitoring is to maintain separate cumulative sum charts over time for each region. A drawback of this approach is that it does not account for the possibility of clusters occurring on larger spatial scales. In this paper, we describe how monitoring may be carried out for neighborhoods constructed around each sub region. Separate charts may be kept for each sub region and its surrounding neighborhood. However, adjusting cusum thresholds for the number of sub regions is conservative, as nearby regions will have correlated charts. Here, these correlations are accounted for; an adjustment for the number of effectively independent charts is made using the theory of smoothed Gaussian random fields, and the approach is evaluated.

15. An Empirical Investigation of the Pareto Principle in the Supermarket Industry Stores.

Research Team: Jeremy Campbell (undergraduate student, University at Buffalo), Dinesh Kumar Gauri (PhD student, University at Buffalo), Debabrata Talukdar (University at Buffalo) and Minakshi Trivedi (University at Buffalo).

Details: The conventional wisdom of the Pareto Principle, more commonly known as the “80/20 rule”, has seen applications across a variety of contexts. Our study focuses on a market segmentation aspect which holds that about 80% of a firm’s sales come from only about 20% of its customers. Such “concentration” of sales revenue from a relatively small group of customers has obvious strategic implications for firms’ decisions regarding target marketing and customer service resource allocation. In this study, we conduct a detailed empirical investigation to investigate the robustness of the 80/20 rule at various levels of aggregation: (1) store level; (2) individual product category levels within a store; and (3) individual leading brands across various product categories within a store.

16. An Integrated Model to Explain Customer Relationships.

Research Team: Sekar Raju (University at Buffalo), and Professor H. Rao Unnava (Ohio State University).

Details: The marketing literature has focused a lot on understanding the role satisfaction plays in building brand loyalty and ensuring repeat purchases. However, data from the field suggests that satisfaction surveys do not do a good job in explaining repeat purchase or the likelihood of customer defection. This research project attempts to integrate other components that affect the decision to continue a relationship with a store/firm. Specifically, we plan to integrate satisfaction with sunk costs and interdependencies between the customer and the store/firm. This model of customer relationship is expected to predict customer loyalty and repeat purchases better than existing models.

17. Objective vs. Subjective Choice Variety: Have Traditional Objective Measures Overestimated (or Underestimated) Choice Variety?

Research Team: Kalpesh Desai (University at Binghamton) and Minakshi Trivedi (University at Buffalo). Authors listed alphabetically.

Details: The past two decades have seen the development of a rich stream of literature in variety seeking in both behavioral as well as modeling domains. Both sub-streams have adopted an “outside-in” perspective to measuring choice variety. That is, using a researcher–defined measure of choice variety, consumers are classified as more vs. less variety seeking based on their purchase history. It is not clear, however, if the consumers choosing the alternatives “subjectively” see these alternatives as different. Even though few studies have used subjective measures of choice variety, no concurrence with objective measures has been sought. The primary objective of this research, then, is to reconcile these two measures of choice variety and address the following fundamental question: Have traditional objective measures of choice variety overestimated or underestimated the choice variety for high variety seekers? Discrepancy between the objective and subjective measures of choice variety hold important implications for market structure differences between the high and low variety seeking segments that prior research has not examined. In addition, the discrepancy also holds important implications for the extent of stimulation provided by brand switching and the role of assortment variety in influencing choice of high variety seekers.

18. Effect of Store Brand Patronage on Store Patronage.

Research Team: K. Sudhir (Yale University) and Debabrata Talukdar (University at Buffalo). Authors listed alphabetically.

Details: We investigate the relationship between a household's store brand patronage and store patronage through its impact on store revenues and profits. The nature of the relationship will help answer the question: Do store brands contribute to greater store differentiation or to greater price sensitivity in the market? Our initial analyses (published in Review of Industrial Organization, 2004) show support for the store differentiation argument. We are currently working on a more comprehensive analysis.

19. “Extreme” Cherry Picking Behavior: Prevalence, Determinants and Impact on Retailers’ Profit

Research Team: Dinesh Kumar Gauri (PhD student, University at Buffalo) and Debabrata Talukdar (University at Buffalo). Authors listed alphabetically.

Details: In this study, the focus is on that segment of consumers who visits a retailer to buy just the deeply discounted items (viz., those items that are put on ‘loss leader’ pricing by the retailer), there by essentially generating negative profit contribution for the retailer. Since loss leader pricing strategy is a very popular promotion strategy for the grocery retailers, it becomes very important for research to focus on the following issues: (1) Estimate the size of the extreme cherry picking segment across different stores in different market contexts; (2) Identify the key market characteristic drivers of such segment size by performing an aggregate level analysis at the store level; (3) Identify the key consumer characteristic drivers of the extreme cherry picking behavior by performing a disaggregate level analysis at the consumer level; and (4) Estimate the incremental net positive impact, if any, of loss leader pricing strategy on the bottom-line of the retailer, which is likely to be the most critical information of interest to a retailer.

20. Evaluating Spatial Performance of a Supermarket Store: Insights and Implications.

Research Team: Andrei Strijnev (University of Texas, Dallas) and Debabrata Talukdar (University at Buffalo). Authors listed alphabetically.

Details: In this research, we seek to address the issue of how a retail grocery store could decide on setting “efficient” targets to evaluate its spatial sales performance. The wide prevalence of “frequent shopper” loyalty card programs by grocery retail chains means that most grocery retail stores collect data that can be used to trace back the spatial distribution of its sales. While such data will thus allow a store to evaluate its spatial sales performance over time, it does not provide any insight into how the store’s spatial sale performance compares to total sales potential in a given spatial unit. The motivation and contribution of this research is to develop a decision model that can provide such insight to store managers. The methodology used will be Stochastic Frontier Analysis for efficient performance evaluation.

21. “Timing” the Market for Frequently Purchased Goods: Consumers’ Performance and Its Determinants.

Research Team: Amresh Kumar (PhD student, University at Buffalo) and Debabrata Talukdar (University at Buffalo). Authors listed alphabetically.

Details: The backdrop for this study is the widely prevalent within-store inter-temporal price variation in markets of frequently purchased goods. Naturally, an interesting question in these markets is whether consumers form price expectations and “lie in wait” to take advantage of the inter-temporal price variations. Past research has indeed found evidence of consumers strategically shifting their purchase decisions (timing and quantity) to take such an advantage. The evidence has come in the form of purchase acceleration (or, “stockpiling”) and deceleration behaviors at individual consumer level, and in the form of pre-promotion and post-promotion dips in sales at store level. However, there has been hardly any systematic study to investigate the revealed effectiveness of such strategic purchase behaviors on the part of consumers in essentially “timing” the market. The insights from such a study hold considerable implications for retailers pricing strategy. The goal of this research is to undertake such a study, which is similar to studies for evaluating performance of “market timing strategy” in financial equity markets.

22. Opportunity Costs of SKU Reductions for Grocery Supermarkets: Insights and Implications for Retailers’ Product Assortment.

Research Team: Amresh Kumar (PhD Student, University at Buffalo) and Debabrata Talukdar (University at Buffalo). Authors listed alphabetically.

Details: Since the 1990s, there has been a significant proliferation in SKUs in supermarkets as both manufacturers and retailers saw it as a strategic way for increasing respective market shares. However, in recent years, the higher costs of carrying a large assortment of SKUs as well as competition from lower cost, lower assortment-carrying discount retailers have led retailers to look for more efficient assortment strategies by eliminating low-selling SKUs. A key input to formulating such strategies by retailers obviously hinges on their ability to have a good understanding of opportunity costs of SKU reductions. The goal of this study is to develop a heuristic approach to empirically estimate such opportunity costs based on the typical scanner data available to a retailer from loyalty card programs, and without requiring a priori data from actual reductions of SKUs under analysis. The opportunity cost estimate will take into account not only the direct cost from lost profit that is directly being generated by a SKU, but also the indirect cost from its elimination.

23. Benchmarking Retail Store Performances Based on “Efficient” Rather than “Average” Performance.

Research Team: K. Sudhir (Yale University) and Debabrata Talukdar (University at Buffalo). Authors listed alphabetically.

Details: Companies set performance targets for many types of employees (e.g., sales force, store managers, hedge fund/mutual fund managers and production managers). In this research project, we seek to address the issue of how a retail chain should decide on performance targets for store managers operating in markets with diverse store characteristics (demographics, competition, store size, etc.). A common approach is to use a regression model where the performance is explained using explanatory variables such as demographic, competitive and other store characteristics. The fitted values from the regression are then used to set performance targets for store managers. But benchmarking obtained through such analysis produces performance objectives based on the “average” performance of their employees, and do not provide incentives for employees to excel. Management would prefer to infer the “efficient frontier,” the best possible performance given the store characteristics, and then use that for benchmarking. Intuitively, we can get such benchmarks by fitting a regression-type model only on stores that are operating with the greatest efficiency. Methodologically, this will require using a linear programming based approach called “Data Envelopment Analysis”.

24. Determinants of New Brand Success.

Research Team: Ram Bezawada (University at Buffalo), Vijay Ganesh Hariharan (PhD Student, University at Buffalo) and Debabrata Talukdar (University at Buffalo). Authors listed alphabetically.

Details: What determines the relative success of new brand introductions? What determines the relative propensity on the part of consumers to try and repeat purchase new brands? Here, at the aggregate level, we will look across categories to see how competitive structure, nature of extension (line or category), the relative market position of the introducing firm, etc affect the new brand market share. At the disaggregate level, we will look across categories to see how consumer differences in category consumption patterns, purchase of other brands of the introducing firm in the same or different category (kind of “firm/brand equity”), etc. affect who adopts the new brand more.

25. Optimal Pricing Policies under Demand Dependencies across Categories.

Research Team: Dinesh Kumar Gauri (PhD student, University at Buffalo) and Debabrata Talukdar (University at Buffalo). Authors listed alphabetically.

Details: The objective of this is study to investigate the interdependencies in consumer demand across product categories and develop retailer pricing strategies under circumstances where demand for two or more product categories may be correlated. Our approach would be to estimate a series of demand functions that explicitly account for the cross-category dependence in household purchase behavior. The demand models would allow for heterogeneity across households due to observed (such as demographics) as well as unobserved factors. The result from the multi-category demand models would be taken as an input to set up a pricing optimization problem for the retailer. In particular, we would extend the standard category profit maximization problem to a multi-category setting to develop a normative pricing policy for the retailers. Our hope is that the models developed in the study can be used as a decision support system for retailers equipped with frequent shopper databases for better decision making.

26. Investigating the Inter Temporal Determinants of Store Traffic Performance using a Dynamic Linear Model.

Research Team: Rajesh K. Shah (University at Buffalo) and Ram Bezawada (University at Buffalo).

Details: Store traffic is the most visible performance measure for store managers as it directly relates to sales response and other marketing mix strategies. Therefore, store managers invest in marketing programs that lead to enhanced store traffic. Thus, the ability of a store to attract customers is an important attribute that can be used as a strategic variable to compete effectively. Moreover, a reliable estimate of store count will help the store to improve the efficiency of its operations and also manufacturers in better directing their promotional effort. Given the above, store traffic is one of the key variables for retailers around which other retail activities are usually centered. Although, issues relating to store traffic have been researched in the marketing literature, they have not been dealt with in a comprehensive manner. For instance, store traffic usually has been analyzed as an outcome of other retail specific activities such as the impact of store level promotions. Examples of the above include the influence of loss-leader items (Walters and MacKenzie 1988), feature promotions (Mulhern and Leone 1990, Bodapati 1999), grocery advertisements and store flyers (Urbany et al. 2000, Gijsbrechts et al. 2003) on store traffic. Moreover, these studies have not considered the dynamic nature of store traffic. Retail strategies with respect to key factors such as store traffic are set over longer time horizons (e.g., AMR 2000). Therefore, the inter-temporal nature of such factors and the manner in which they influence store traffic needs to be explored. In this paper, we address the above issue using a dynamic linear model to account for the evolution of store traffic. We use Bayesian parameter estimation and model selection criteria to model inter temporal effects of various retailer strategies. In addition to marketing mix variables such as features, loss leaders and prices; we also consider how other store specific characteristics such as assortment levels, produce quality, organic products influence store traffic. We further include the store descriptors like physical characteristics (e.g. size, formats, personnel) and non-traditional product offerings (e.g. presence of a bank, gas stations, café). Our results will better enable retail managers to understand multilevel-determinants of store traffic and develop strategies to monitor and enhance store traffic.

27. An Empirical Investigation of the Effect of the Time of Shopping on the Inter-Purchase Time of Grocery Products.

Research Team: Kalpesh Desai (University at Binghamton) and Dinesh Kumar Gauri (PhD student, University at Buffalo). Authors listed alphabetically.

Details: To ensure steady business, retailers prefer their shoppers to shop regularly for frequently purchased products i.e., they prefer shoppers to maintain consistency in their inter-purchase time of such products. Memory literature has found evidence of the time of day effect on recall of products e.g., orange juice (beer) is likely to be recalled more during morning vs. evening (evening vs. morning) (Unnava et al. 2006). However, there is no empirical investigation if similar effects occur in the purchase of grocery products. In this research we examine the inter-purchase timing behavior of consumers across various product categories and hypothesize a three-way interaction between shopper’s basket size X shopper’s “typical” shopping time X time-related products (e.g., orange juice is a morning product and beer is an evening product). Specifically, we posit there will be greater consistency in the inter-purchase time of “matched” (mismatched) products for small (large) basket shoppers. Matched products refer to morning products purchased in morning and evening products purchased in evening. We also expect inter-purchase time consistency for time independent products (e.g., bathroom tissues, laundry detergent) to be lower among small vs. large basket shoppers. In addition to the steady business managerial implication, findings from this research will help in selecting appropriate categories for promotions based on the time of the day and in micro-targeting promotions to shoppers based on their basket size and time of shopping.