Gary J. Russell- PhD, University of Chicago
- Professor at University of Iowa
Gary J. Russell
- PhD, University of Chicago
- Professor at University of Iowa
About
44
Publications
41,534
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
4,564
Citations
Introduction
Gary J. Russell is Henry B. Tippie Research Professor of Marketing at the Tippie College of Business, University of Iowa. His primary research interest is the application of scanner-data-based choice models to substantive marketing issues. His work has addressed sales response to advertising, market structure definition, brand equity measurement, and brand price competition. Recent work is concerned with multiple category promotion response, fusion methodologies for linking panel and store-level scanner data, and spatial aspects of choice. His 1989 JMR article on the latent class logit model with Wagner Kamakura (Rice University) is ranked 30th on a list of the 100 most influential articles in marketing science (2014 survey reported in International Journal of Research in Marketing).
Current institution
Publications
Publications (44)
The history of choice modeling in marketing science is a meandering path, informed by work in psychology, economics and statistics. Early work by Thurstone (1927), Luce (1959), Tversky (1972) and McFadden (1980) provided the theoretical foundations for the analysis of scanner panel data using the multinomial logit model. Advances in computational p...
Consumers in grocery retailing commonly buy bundles of products to accommodate current and future consumption. When all products in a particular bundle share common attributes (and are selected from the same product category), the consumer is said to assemble an assortment. This research examines the impact of assortment variety on the assortment c...
Market basket choice is a decision process in which a consumer selects items from a number of product categories on the same shopping trip. The key feature of market basket choice is the interdependence in demand relationships across the items in the final basket. This research develops a new approach to the specification of market basket models th...
Product recommendation models are key tools in customer relationship management (CRM). This study develops a product recommendation model based upon the principle that customer preference similarity stemming from prior response behavior is a key element in predicting current product purchase. The proposed recommendation model is dependent upon two...
Purpose
This paper aims to analyze variations in the parameters of the market share–rank power law across consumer packaged goods (CPG) categories.
Design/methodology/approach
The authors use a two-level hierarchical linear model to examine the relationships between category-level variables and the parameters of the market share–rank power law in...
This study examines the relationship between market share and its rank for a comprehensive set of product categories and subcategories in the U.S consumer packaged goods industry. Based on prior research, we expect that the slope of the market share-rank relationship will be more consistent with the Zipf distribution at the higher levels of aggrega...
Emerging Trends in Product Bundling: Investigating Consumer Choice and Firm Behavior Bundling is the practice of selling two or more products together, often at a discounted price. In this article, we extend the concept of bundling to a wide variety of choice settings. We argue that bundle choice covers consumer decision scenarios which differ with...
Most models for the diffusion of innovation are deeply rooted in the work of Bass (1969). The Bass model, however, assumes all potential customers are linked to all other customers in a social network. In this research, we develop the Network-Based Bass model (NBB), a generalization of the Bass model that incorporates observed individual-level soci...
Understanding customer preference for bundles is key to the success of this widely-used segmented pricing strategy. In addition to conjoint analysis, researchers and practitioners often use a " build-your-own-bundle " or configuration approach to understand customer preferences for bundles. In a configuration study, subjects are asked to create an...
Sentient Machine Research, a Dutch data mining company, for providing access to the data used in this research. A SPATIAL CHOICE MODEL FOR PRODUCT RECOMMENDATIONS Product recommendation models are key tools in customer relationship management (CRM). This study develops a product recommendation model based upon the principle that customer preference...
Database marketers often select households for individual marketing contacts using information on past purchase behavior. One of the most common methods, known as RFM variables approach, ranks households according to three criteria: the recency of the latest purchase event, the long-run frequency of purchases, and the cumulative dollar expenditure....
providing access to the data used in this research. Profiling the Reference Price Consumer Researchers in marketing have devoted considerable attention to understanding how price impacts the purchase decision. Some individuals (termed memory-based reference price consumers) take into account price expectations developed from past purchase behavior...
One of the key problems in database marketing is the identication and proling of households who are most likely to be interested in a particular product or service. Principal component analysis (PCA) of customer background information followed by logistic regression analysis of response behavior is commonly used by database mar- keters. In this pap...
Marketing science models typically assume that responses of one entity (firm or consumer) are unrelated to responses of other entities. In contrast, models constructed using tools from spatial statistics allow for cross-sectional and longitudinal correlations among responses to be explicitly modeled by locating entities on some type of map. By gene...
We empirically test existing theories on the provision of public goods, in particular air quality, using data on sulfur dioxide (SO2) concentrations from the Global Environment Monitoring Projects for 107 cities in 42 countries from 1971 to 1996. The results are as follows: First, we provide additional support for the claim that the degree of democ...
One of the key problems in database marketing is the identification and profiling of households who are most likely to be interested in a particular product or service.
Direct Marketing Educational Foundation for providing access to the data used in this
We consider customer influences on market structure, arguing that market structure should explain the extent to which any given set of market offerings are substitutes or complements. We describe recent additions to the market structure analysis literature and identify promising directions for new research in market structure analysis. Impressive a...
In many purchase environments, consumers use information from a number of product categories prior to making a decision. These purchase situations create dependencies in choice outcomes across categories. As such, these decision problems cannot be easily modeled using the single-category, single-choice paradigm commonly used by researchers in marke...
The authors derive a theoretical relationship between the aggregate market share elasticity matrix and the aggregate brand switching matrix on the basis of a logit model of heterogeneous consumers choosing among competing brands in a product class. Aggregate cross-elasticities are shown to be proportional (through a single scaling constant) to thei...
The authors derive a theoretical relationship between the aggregate market share elasticity matrix and the aggregate brand switching matrix on the basis of a logit model of heterogeneous consumers choosing among competing brands in a product class. Aggregate cross-elasticities are shown to be proportional (through a single scaling constant) to thei...
Household basket data contain important information about the structure of brand preferences both within and across product categories. This research exploits the information in long-run basket summary data to segment consumers with respect to brand preferences. The approach provides insights into the competitive structure of brands within each pro...
Multiple category choice is a decision process in which an individualselects a number of goods, all of which are nonsubstitutable with respect toconsumption. Choices can be made either simultaneously or sequentially. Thekey feature of multiple category choice is the treatment of the choices asinterrelated because each item in the final collection o...
Recent advances in data gathering through checkout scanners have produced vast amounts of data on the actual behavior of consumers in the marketplace, creating new opportunities for managers and researchers to understand competition and consumers' response to the marketing mix. Previous analyses of this data in the literature have focused either at...
Recent advances in data gathering through checkout scanners have produced vast amounts of data on the actual behavior of consumers in the marketplace, creating new opportunities for managers and researchers to understand competition and consumers’ response to the marketing mix. Previous analyses of this data in the literature have focused either at...
The authors develop an approach to decompose a market-level matrix of own- and cross-price elasticities to reveal potentially overlapping preference segments. The approach is grounded on the premise that markets may be represented by a parsimonious number of relatively homogeneous segments. Market-level elasticities are expressed as functions of se...
Using actual consumer choice data from a single-source scanner panel, we construct two measures of brand value which capture different aspects of brand equity. Brand Value measures perceived quality, the value assigned by consumers to the brand, after discounting for current price and recent advertising exposures. Brand Intangible Value isolates th...
This paper develops the Latent Symmetric Elasticity Structure (LSES), a market share price elasticity model which allows elasticities to be decomposed into two components: a symmetric substitution index revealing the strength of competition between brand pairs, and a brand-specific coefficient revealing the overall impact of a brand on its competit...
Marketing scholars commonly characterize market structure by studying the patterns of substitution implied by brand switching. Though the approach is useful, it typically ignores the destabilizing role of marketing variables (e.g., price) in switching behavior. The authors propose a flexible choice model that partitions the market into consumer seg...
Marketing scholars commonly characterize market structure by studying the patterns of substitution implied by brand switching. Though the approach is useful, it typically ignores the destabilizing role of marketing variables (e.g., price) in switching behavior. The authors propose a flexible choice model that partitions the market into consumer seg...
Data interval bias, the biased estimation of advertising carryover in aggregate data, can be viewed as a misinterpretation of the aggregate advertising-sales relationship due to missing micro advertising data. This paper argues that if the researcher does not explicitly model the firm's advertising decisions, he will incorrectly interpolate the mis...
Though considerable attention has been given to market structure, little research has been done on the relationship between market structure and elasticity structure. The authors develop and partially test the aggregate constant ratio elasticity pattern (ACREP), a parsimonious marketing mix elasticity model that describes the elasticity structure o...
Though considerable attention has been given to market structure, little research has been done on the relationship between market structure and elasticity structure. The authors develop and partially test the aggregate constant ratio elasticity pattern (ACREP), a parsimonious marketing mix elasticity model that describes the elasticity structure o...
Though the PIMS data have been used extensively to relate profitability both to specific firm marketing actions and to market conditions, no attempt has been made to estimate directly the magnitudes of marketing mix elasticities. The authors propose and evaluate a simple cross-sectional time series regression method that recovers unbiased estimates...
Though the PIMS data have been used extensively to relate profitability both to specific firm marketing actions and to market conditions, no attempt has been made to estimate directly the magnitudes of marketing mix elasticities. The authors propose and evaluate a simple cross-sectional time series regression method that recovers unbiased estimates...
Lists of nutrition information posted in supermarkets were designed to reduce the information-processing costs of comparing alternative foods. In Experiment 1, lists of vitamins and minerals increased nutrition knowledge but had no influence on actual purchases. In Experiment 2, a list of added sugar—a negative component of food—increased the marke...
The factors that influence people to come for psychiatric treatment are not well known. This report comes from a long-term prospective study of the 1966–1967 first-grade population of children and families in Woodlawn, a poor, black, urban community on Chicago's South Side. These children, now adolescents, were recently reassessed in a 10-year foll...
Fred Feinberg is Michael R. and Mary Kay Hallman Fellow and Associate Professor of Marketing at the University of Michigan Business School. Gary J. Russell is Henry B. Tippie Research Fellow and Professor of Marketing at the Henry B. Tippie College of Business, University of Iowa. The names of the authors are listed alphabetically; both contributed...
Ann Petersen, Ph. D. is Principal of the Halley Group. Gary J. Russell is Henry B. Tippie Research Fellow and Professor of Marketing at the Tippie College of Business, University of Iowa. Doyle Weiss is Professor of Marketing at the Tippie College of Business, University of Iowa. This article is based upon the first authors'doctoral research at the...
Choice decisions in market basket selection are interrelated because demand patterns are complex. This research develops a SKU-level multiple-item choice model, based upon the assumption that the utility of purchase of one item in a market basket depends upon the set of items already selected. The resulting model is very flexible, allowing for a la...
Questions
Question (1)
The control function (CF) approach to endogeneity correction is a fully general technique that can be used for nonlinear as well as linear models. Suppose that we use CF to control for (say) price endogeneity in the logit choice model. Once the CF variable is included, we will obtain consistent estimate of all parameters (including price).
HOWEVER ... Choice models, such as logit, have the property that coefficients are scaled by the error standard deviation. This implies, in particular, that the model coefficients will be bigger when the model fits the data better. By including a CF variable, we alter the model fit and thus rescale all model parameters. For forecasting purposes, then, we cannot simply set the CF coefficient to zero and then compute forecasts from the remaining variables. (The scaling is different in a model with the CF variable and a model without the CF variable.)
So ... the question is what to do. Should we simply use the logit model "as is" by including the CF variable in making forecasts? Or should we delete the CF variable, rescale the coefficients, and then make forecasts? In the latter case, how do we figure out what the appropriate scaling coefficient should be?
AN EXAMPLE ... The attached article uses the CF approach for logit choice models and forecasts choice probabilities by leaving the CF variable in the logit equation. Is this the correct approach to using the CF approach in choice models?
