Thought Leadership

Want to learn the nitty-gritty of what we do? Below are articles, conference decks, and whitepapers written by members of our senior leadership team. Learn more about the following topics:

Maximum Difference Scaling (MaxDiff)

What's your preference? Asking survey respondents about their preferences creates new scaling decisions. (1.15 MB)
Published: 2004
Author: Steven H. Cohen and Bryan Orme

When it comes to scaling multiple items, researchers have several options. This article compares maximum difference scaling against monadic ratings and paired comparisons. Among other benefits, it offers improved measures of discrimination across items over rating scales. This paper won the David K. Hardin Award for Best Paper in 2004 Marketing Research Magazine.


Maximum Difference Scaling: Improved Measures of Importance and Preference for Segmentation. (334.65 kB)
Published: 2003
Author: Steven H. Cohen

Maximum Difference scaling (MaxDiff), a technique for measuring the importance or preference of multiple items, is shown to provide results that have greater between-item and between-respondent discrimination, and greater predictive accuracy than either monadic ratings or paired comparisons. Steve Cohen describes the methodology and presents results for a methodological study comparing MaxDiff measurement with monadic ratings and paired comparisons, and also a case study focusing on using MaxDiff for segmentation work. Steve won the "best presentation" award with this paper at the 2003 Sawtooth Software Conference.


Measuring preference for product benefits across countries: Overcoming scale usage bias with Maximum Difference Scaling. (276.88 kB)
Published: 2003
Author: Steven H. Cohen and Leopoldo Neira

This paper briefly reviews the standard practices of benefit measurement and benefit segmentation and, along the way, points out their deficiencies. The authors then introduce Maximum Difference scaling, a method they believe is a much more powerful method for measuring benefit importance. An example is provided of how both the traditional and the newer methods were used in a cross-national consumer study of coffee drinkers and how the results compare. This paper is the winner of the John and Mary Goodyear Award for the Best International Paper presented at ESOMAR Conferences in 2003.


Renewing market segmentation: Some new tools to correct old problems. (93.27 kB)
Published: 2002
Author: Steven H. Cohen and Paul Markowitz

Nominated for Best Methodological Paper at the 2002 ESOMAR Congress, this paper critically examines the usual and standard tools and methods of benefit segmentation as practiced by most researchers and suggests new ways of renewing market segmentation through methods of measuring benefit importance and performing segmentation analysis on the resulting data. The authors introduce Maximum Difference Scaling, a more powerful method for measuring benefit importance that is scale-free and thus very applicable to international segmentation research. The paper describes how Maximum Difference Scaling can be combined with Latent Class Analysis to obtain international benefit segments.


Segmentation & Latent Class Models

Maximum Difference Scaling: Improved Measures of Importance and Preference for Segmentation. (334.65 kB)

Published: 2003
Author: Steven H. Cohen

Maximum Difference scaling (MaxDiff), a technique for measuring the importance or preference of multiple items, is shown to provide results that have greater between-item and between-respondent discrimination, and greater predictive accuracy than either monadic ratings or paired comparisons. Steve Cohen describes the methodology and presents results for a methodological study comparing MaxDiff measurement with monadic ratings and paired comparisons, and also a case study focusing on using MaxDiff for segmentation work. Steve won the "best presentation" award with this paper at the 2003 Sawtooth Software Conference.


Measuring preference for product benefits across countries: Overcoming scale usage bias with Maximum Difference Scaling. (276.88 kB)
Published: 2003
Author: Steven H. Cohen and Leopoldo Neira

This paper briefly reviews the standard practices of benefit measurement and benefit segmentation and, along the way, points out their deficiencies. The authors then introduce Maximum Difference scaling, a method they believe is a much more powerful method for measuring benefit importance. An example is provided of how both the traditional and the newer methods were used in a cross-national consumer study of coffee drinkers and how the results compare. This paper is the winner of the John and Mary Goodyear Award for the Best International Paper presented at ESOMAR Conferences in 2003.


Renewing market segmentation: Some new tools to correct old problems. (93.27 kB)
Published: 2002
Author: Steven H. Cohen and Paul Markowitz

Nominated for Best Methodological Paper at the 2002 ESOMAR Congress, this paper critically examines the usual and standard tools and methods of benefit segmentation as practiced by most researchers and suggests new ways of renewing market segmentation through methods of measuring benefit importance and performing segmentation analysis on the resulting data. The authors introduce Maximum Difference Scaling, a more powerful method for measuring benefit importance that is scale-free and thus very applicable to international segmentation research. The paper describes how Maximum Difference Scaling can be combined with Latent Class Analysis to obtain international benefit segments.


Latent Segmentation Models: New tools to assist researchers in market segmentation. (1.16 MB)
Published: 1998
Author: Steven H. Cohen and Venkatram Ramaswamy

Market segmentation remains one of the most powerful marketing ideas. Since its formal introduction in the '50s, its use for customer understanding, product development and marketing strategy has grown. The development of analytic methods for segmenting markets, however, has lagged their need in business applications. The authors see the next wave as a suite of analytic procedures called latent segmentation models.


Joint Segmentation on Distinct Interdependent Bases with Categorical Data. (1.54 MB)
Published: 1996
Author: Venkatram Ramaswamy, Rabikar Chatterjee, and Steven H. Cohen

The authors discuss a latent class framework for market segmentation with categorical data on two conceptually distinct but possibly interdependent bases for segmentation (e.g., benefits sought and usage of products and services). The authors present an empirical application, using pick-any data collected by a regional bank on two popular, conceptually appealing, and interdependent bases for segmenting customers of financial services-benefits (i.e., desired financial goals) and product usage (of an array of banking services). A comparative evaluation of the approach on synthetic data demonstrates the ability of the modeling framework to detect and estimate the interdependence structure underlying the two segmentation bases and thereby provide more accurate segmentation than "traditional" (single-basis) latent segmentation methods.


Market segmentation with Choice-based Conjoint Analysis. (1.11 MB)
Published: 1995
Author: Wayne S. Desarbo, Venkatram Ramaswamy, and Steven H. Cohen

We discuss a method for market segmentation with choice-based conjoint models. This method determines the number of market segments, the size of each market segment, and the values of segment-level conjoint part-worths using commonly collected conjoint choice data. A major advantage of the proposed method is that current (incomplete) data collection approaches for choice-based conjoint analysis can still be used for market segmentation without having to collect additional data. We illustrate the proposed method using commercial conjoint choice data gathered in a new concept test for a major consumer packaged goods company. We also compare the proposed method with an a priori segmentation approach based on individual choice frequencies.


Conjoint Analysis, Choice Models & Choice-Based
Conjoint Analysis (CBCA)

Have it Your Way: Menu-based conjoint analysis helps marketers understand mass customization. (1.09 MB)
Published: 2007
Author: Steven H. Cohen and John C. Liechty

How must researchers craft studies and analyze the data to get a deeper grasp of mass customization? How must researchers investigate picking from a menu, and how must they examine the results to best expose the behavioral underpinnings of the build-your-own situation? One might think that traditional conjoint analysis is appropriate and effective for understanding how people want to construct product bundles, but the menu situation's additional complexity makes that approach entirely inadequate. Enter menu-based conjoint analysis – expressly designed for handling a build-your-own, select-from-a-menu, mass customization situation.


When Consumers Go beyond Choice: Models for Trade-up and Change in Consideration Set. (268.39 kB)
Published: 2007
Author: Mark Garratt and Greg Allenby

Presented at the 2007 ART Forum Conference in Santa Fe, New Mexico, this conference deck explains trade-up decisions made by consumers and introduces the presenters' trade-up model that measures this effect.


A Choice Model for Packaged Goods: Dealing with Discrete Quantities and Quantity Discounts. (232.46 kB)
Published: 2004
Author: Greg Allenby, Tom Shively, Sha Yang, and Mark Garratt

In this paper, we provide an economic model of demand for substitute brands that is flexible, parsimonious, and easy to implement. The methodology is demonstrated with a scanner panel data set of light-beer purchases. The model is used to explore the effects of price promotions on primary and secondary demand, and the utility of product assortment.


Developing an Optimal Product Line Using Consumer Preferences and Cost Data. (278.77 kB)
Published: 2003
Author: Steven H. Cohen and Susan Stoev

Abstract: Much of the activity of market researchers focuses on understanding the effect of new products on the firm. By researching product acceptance before introduction, manufacturers recognize that careful attention to consumer wants and needs will increase their likelihood of success. This paper provides a case study of the challenges that faced both a client firm (Eastman Kodak Company) and its research partner (SHC & Associates) in developing research to study consumer acceptance of potential entrants in four new product lines being contemplated by Kodak. The paper first addresses the challenges faced by the internal researcher in drawing out the needs of product management, having them accept the possibilities and limitations of consumer research, and developing an understanding of the study findings and uses of the research when the study is complete. Secondly the paper addresses the challenges faced by the research agency in developing a research design and program that fits the needs of the client, implementing that design, and developing findings that can be easily understood and implemented.


Choice-Menus for Mass Customization: An Experimental Approach for Analyzing Customer Demand With an Application to a Web-based Information Service (182.81 kB)
Published: 2001
Author: John Liechty, Venkatram Ramaswamy, and Steven H. Cohen

Companies are increasingly engaging in mass customization and offering consumers a "choiceboard" (or a menu of choices) of various features and options for configuring their own products and services. The authors discuss the use of experimental choice menus for assessing customers' preferences and price sensitivities for the variety of features and options that might be offered by a firm in its choiceboard.


Perfect Union: CBCA marries the best of conjoint and discrete choice models. (941.76 kB)
Published: 1997
Author: Steven H. Cohen

This article reports on discrete choice models as utilized by choice-based conjoint analysis (CBCA). The nature of the discrete choice model involves statistical methods for choice response analysis while conjoint analysis requires a research technique for a systematic collection of data on choices using experimental design. Both assist managers and researchers in reading drivers of choice. During the early 1970's, conjoint analysis was used to interpret consumer preferences in a wide range of market situations. At about the same time, discrete choice models were introduced to explain the behavior of consumers. Characteristics of both were merged to form CBCA, which subdues deficiencies and offers more choices from the set of product alternatives.


Forecasting the Market Potential for New Providers of Local Telephone Services to Business Customers. (47.4 kB)
Published: 1996
Author: Steven H. Cohen, Donald A. Anderson, and Jill Hesser

At the time this paper was written, deregulation was creating competition in the communications industry, allowing telephone service providers, media companies and other firms to offer everything from local telephone services to video-on-demand. Thus, predicting preference and likely market share for new competitors was a critical business issue. This case study is based on research conducted by U S WEST and Time Warner Communications to quantify the potential for medium and large corporations and government agencies to switch their business from the incumbent local telephone provider to companies offering similar services. We describe our approach to predicting preferences and likely market share using a designed discrete choice experiment. We also show how customer preferences were combined with managerial judgments of the potential behaviors of likely competitors to predict share gains and losses using a dynamic diffusion model.


Market segmentation with Choice-based Conjoint Analysis. (1.11 MB)
Published: 1995
Author: Wayne S. Desarbo, Venkatram Ramaswamy, and Steven H. Cohen

We discuss a method for market segmentation with choice-based conjoint models. This method determines the number of market segments, the size of each market segment, and the values of segment-level conjoint part-worths using commonly collected conjoint choice data. A major advantage of the proposed method is that current (incomplete) data collection approaches for choice-based conjoint analysis can still be used for market segmentation without having to collect additional data. We illustrate the proposed method using commercial conjoint choice data gathered in a new concept test for a major consumer packaged goods company. We also compare the proposed method with an a priori segmentation approach based on individual choice frequencies.


Efficient Experimental Design with Market Research Applications (1.1 MB)
Published: 1994
Author: Warren F. Kuhfeld, Randall D. Tobias, and Mark Garratt

The authors suggest the use of D-efficient experimental designs for conjoint and discrete-choice studies, discussing orthogonal arrays, nonorthogonal designs, relative efficiency, and nonorthogonal design algorithms. They construct designs for a choice study with asymmetry and interactions and for a conjoint study with blocks and aggregate interactions.


Menu Choice, Bundling & Mass Customization

Have it Your Way: Menu-based conjoint analysis helps marketers understand mass customization. (1.09 MB)
Published: 2007
Author: Steven H. Cohen and John C. Liechty

How must researchers craft studies and analyze the data to get a deeper grasp of mass customization? How must researchers investigate picking from a menu, and how must they examine the results to best expose the behavioral underpinnings of the build-your-own situation? One might think that traditional conjoint analysis is appropriate and effective for understanding how people want to construct product bundles, but the menu situation's additional complexity makes that approach entirely inadequate. Enter menu-based conjoint analysis – expressly designed for handling a build-your-own, select-from-a-menu, mass customization situation.


Developing an Optimal Product Line Using Consumer Preferences and Cost Data. (278.77 kB)
Published: 2003
Author: Steven H. Cohen and Susan Stoev

Much of the activity of market researchers focuses on understanding the effect of new products on the firm. By researching product acceptance before introduction, manufacturers recognize that careful attention to consumer wants and needs will increase their likelihood of success. This paper provides a case study of the challenges that faced both a client firm (Eastman Kodak Company) and its research partner (SHC & Associates) in developing research to study consumer acceptance of potential entrants in four new product lines being contemplated by Kodak. The paper first addresses the challenges faced by the internal researcher in drawing out the needs of product management, having them accept the possibilities and limitations of consumer research, and developing an understanding of the study findings and uses of the research when the study is complete. Secondly the paper addresses the challenges faced by the research agency in developing a research design and program that fits the needs of the client, implementing that design, and developing findings that can be easily understood and implemented.


Choice-Menus for Mass Customization: An Experimental Approach for Analyzing Customer Demand With an Application to a Web-based Information Service (182.81 kB)
Published: 2001
Author: John Liechty, Venkatram Ramaswamy, and Steven H. Cohen

Companies are increasingly engaging in mass customization and offering consumers a "choiceboard" (or a menu of choices) of various features and options for configuring their own products and services. The authors discuss the use of experimental choice menus for assessing customers' preferences and price sensitivities for the variety of features and options that might be offered by a firm in its choiceboard.


Sensory Research

BaSIC (Bayesian Sensory Model Integrated with Characteristics): A Stochastic MDS Model for Sensory Analysis (236.63 kB)
Published: 2011
Author: in4mation insights

Food technologists and perfumers seek to understand how to combine dozens of ingredients and how the blend of these will affect consumers' taste perceptions and product preferences. Increasingly, sensory research has expanded beyond expert panels and chemical analyses to include the analysis of judgments of products by consumers in terms of their preferences and liking. Our solution uses Bayesian estimation methods and is called BaSIC (Bayesian Sensory Model Integrated with Characteristics). The Bayesian statistical framework provides a unique and attractive way of estimating MDS models that addresses many of the challenges of analyzing sensory preference data.


Trade Up & Trade Down Models

A Model for Trade-Up and Change in Considered Brands (189.38 kB)
Published: 2008
Author: Greg Allenby, Mark Garratt, and Peter E. Rossi

A common theme in the marketing literature is the acquisition and retention of customers as they trade-up from inexpensive, introductory offerings to those of higher quality. Standard models of choice, however, apply to narrowly defined categories for which assumptions of near-perfect-substitution are valid. We extend the non-homothetic choice model of Allenby and Rossi (1991) to accommodate effects of advertising, professional recommendation and other factors that facilitate the description and management of trade-up. The model is applied to a national study of an over-the-counter health product.


When Consumers Go beyond Choice: Models for Trade-up and Change in Consideration Set. (268.39 kB)
Published: 2007
Author: Mark Garratt and Greg Allenby

Presented at the 2007 ART Forum Conference in Santa Fe, New Mexico, this conference deck explains trade-up decisions made by consumers and introduces the presenters' trade-up model that measures this effect.


A Choice Model for Packaged Goods: Dealing with Discrete Quantities and Quantity Discounts. (232.46 kB)
Published: 2004
Author: Greg Allenby, Tom Shively, Sha Yang, and Mark Garratt

In this paper, we provide an economic model of demand for substitute brands that is flexible, parsimonious, and easy to implement. The methodology is demonstrated with a scanner panel data set of light-beer purchases. The model is used to explore the effects of price promotions on primary and secondary demand, and the utility of product assortment.


Marketing ROI

From Brilliant to Actionable: It Takes Technical Brilliance and Constant Questioning to Achieve the Truly Actionable in Marketing ROI (2.67 MB)
Published: 2011
Author: Mark Garratt, Rafael Alcaraz, and Steven Cohen

Presented at the 2011 ESOMAR conference, this presentation deck explains the joint effort between Hershey's and in4mation insights to revolutionize how marketing ROI is modeled and how the results of this effort are spread across an organization.


The Next Generation of Marketing Mix Modeling, Analytics, and Optimization (204.69 kB)
Published: 2011
Author: Mark Garratt

in4mation insights provides our clients with a proven, state-of-the-art, next generation approach to Marketing Mix Modeling. Our advanced Bayesian models, industrial-strength hardware/software production facilities, highly-skilled analysts, advanced optimization technologies, and easy-to-use, targeted deliverables are highly superior to those available from traditional suppliers. in4mation insights puts our industry-leading Bayesian technology and strategic thought leadership to work for our clients. The result is simple: we provide more precise and comprehensive category insights that can deliver real differential advantage through increased market share and/or profits.


Bias in the Calculation of Marketing ROI: A Whitepaper (547.89 kB)
Published: 2011
Author: Dr. Sanjib Mohanty and Mark Garratt

By using a common profit term for each element of the marketing mix and by not correctly allocating costs in the profit calculations, the common definition and calculation of ROI will be biased by a large amount. Using the "standard" definition of profit and applying it to each marketing mix element does not enable us to recover the total marginal contribution of advertising and trade from the per unit marginal contributions. The standard profit definition leads to biased ROI because the cost of generating incremental sales is distributed over sales and profits that have nothing to do with those incremental sales. The overestimation is particularly high for advertising, where the share of spending can exceed the contribution by a large factor and may be underestimated for online marketing where the share of spend is often less than the contribution. Correct assignment of per unit costs leads to lower advertising ROIs yet puts more emphasis on establishing the value of the long term return of advertising versus the short term.


Advanced Pricing Models

In Today's Environment, Affordability is More Meaningful Than Elasticity. (1.06 MB)
Published: 2010
Author: Mark Garratt

Price elasticities are the "net" of many different effects driven by all the players in the supply chain: manufacturer, retailer and consumer. Presented at the 2010 Professional Pricing Society Conference Chicago, this conference deck uses new research in consumer behavioral models to show that what we see as price elasticity in aggregate and conjoint data is just as much the consumer decision to buy in the category at all as it is substitution within the category.


When Consumers Go beyond Choice: Models for Trade-up and Change in Consideration Set. (268.39 kB)
Published: 2007
Author: Mark Garratt and Greg Allenby

Presented at the 2007 ART Forum Conference in Santa Fe, New Mexico, this conference deck explains trade-up decisions made by consumers and introduces the presenters' trade-up model that measures this effect.


A Choice Model for Packaged Goods: Dealing with Discrete Quantities and Quantity Discounts. (232.46 kB)
Published: 2004
Author: Greg Allenby, Tom Shively, Sha Yang, and Mark Garratt

In this paper, we provide an economic model of demand for substitute brands that is flexible, parsimonious, and easy to implement. The methodology is demonstrated with a scanner panel data set of light-beer purchases. The model is used to explore the effects of price promotions on primary and secondary demand, and the utility of product assortment.


Other Topics

Exploring the impact of emotions and motivations on hedonics and consumer choice. (847.81 kB)
Published: 2010
Author: Mark Garratt

Presented at the 2010 Institute of Food Technologists (IFT) Food Expo in Chicago, this conference deck provides a review of how consumer insights can be used to develop messaging around the consumer expectations of sensory experiences.


Developing an Optimal Product Line Using Consumer Preferences and Cost Data. (278.77 kB)
Published: 2003
Author: Steven H. Cohen and Susan Stoev

Much of the activity of market researchers focuses on understanding the effect of new products on the firm. By researching product acceptance before introduction, manufacturers recognize that careful attention to consumer wants and needs will increase their likelihood of success. This paper provides a case study of the challenges that faced both a client firm (Eastman Kodak Company) and its research partner (SHC & Associates) in developing research to study consumer acceptance of potential entrants in four new product lines being contemplated by Kodak. The paper first addresses the challenges faced by the internal researcher in drawing out the needs of product management, having them accept the possibilities and limitations of consumer research, and developing an understanding of the study findings and uses of the research when the study is complete. Secondly the paper addresses the challenges faced by the research agency in developing a research design and program that fits the needs of the client, implementing that design, and developing findings that can be easily understood and implemented.


Modeling Variation in Brand Preferences: The Roles of Objective Environment and Motivating Conditions (650.38 kB)
Published: 2002
Author: Sha Yang, Greg Allenby, and Geraldine Fennell

Project designed and comment provided by Mark Garratt on behalf of Miller Brewing Co.


A Unified Imputation Approach for the Treatment and Analysis of Missing Data in Market Research (490.54 kB)
Published: 1999
Author: T.E. Raghunathan, Venkatram Ramaswamy, Steven H. Cohen, and Kerimcan Ozcan

We present a unified approach that integrates multiple imputation with simulation, for the treatment and analysis of missing data in marketing research. It is a generalized approach that can accommodate continuous, categorical or censored data, and is suitable for many common situations encountered in practice. A simulation study with 2000 synthetic data sets demonstrate the robustness and efficiency of MIGS, when compared with alternative approaches. An illustrative commercial application is also provided. We conclude by offering some guidance for dealing with missing data in marketing research.


Efficient Experimental Design with Market Research Applications (1.1 MB)
Published: 1994
Author: Warren F. Kuhfeld, Randall D. Tobias, and Mark Garratt

The authors suggest the use of D-efficient experimental designs for conjoint and discrete-choice studies, discussing orthogonal arrays, nonorthogonal designs, relative efficiency, and nonorthogonal design algorithms. They construct designs for a choice study with asymmetry and interactions and for a conjoint study with blocks and aggregate interactions.