International Choice Modelling Conference, International Choice Modelling Conference 2017

Font Size: 
EFFECTS OF CONFOUNDING CONSUMERS’ CHOICE PROCESS HETEROGENEITY AS TASTE HETEROGENEITY ON THE FIRM DECISION MAKING
Luis Pilli, Joffre Swait, Jose Afonso Mazzon

Last modified: 28 March 2017

Abstract


This project studies the effects on the firm’s performance of misattributing process heterogeneity to tastes as a function of the properties of the firm’s production function, of its monopoly power and of the extent of choice process heterogeneity on the demand side.

In marketing, the identification of consumer response heterogeneity is the foundation for segmentation (where to compete) and positioning (how to compete). This implies that the strategic decisions at the marketing level are outcomes of the policy makers’ understanding about consumers’ preferences that would shape market structure and competition patterns. The misunderstanding of consumers’ preferences and demand structure make firms’ performance vulnerable as the statistics on new product failure clearly demonstrate. Accordingly to Nielsen’s Breakthrough Innovation Report Europe 2014, 76% of new product launches in Western Europe did not achieve 52 weeks of sales and almost 50% did not survive the 26th week.

The basic proposition of this paper is that misattributing process heterogeneity to taste is likely to lead firms to be driven away from profit maximization through: (i) first, a wrong understanding of demand will be expressed via the biased location of the utility parameters, leading to the inference of wrong policy measures. More important, the misattribution is likely to suggest variance in the taste parameters supporting the existence of segments that are not actually present in the marketplace; (ii) second, the firm will make long term commitments based on its understanding of the market structure. This will result in an updated production (and cost) function that supposedly maximizes profit given the industry structure; and (iii) finally, when the actual demand does not confirm the supposed market heterogeneity, the firm will be forced to make short term adjustments of production (and costs), away from the optimal.

To study the effect of misattributing process heterogeneity to preferences on the firm’s strategic decisions, the project unfolds in two steps. In the first step, a demand representation is built from a Monte Carlo simulation based on knowledge of the true data generation process, absent taste heterogeneity. As a result, a true and a misspecified representation of consumers’ preferences and market structure will be plugged into a production function that shall allow for the estimation of the effects of the misattribution of consumers’ response heterogeneity in the firm’s outcomes. The severity of the consequences will be studied as a function of characteristics that may be observed either on the firm side or on the demand side.

Starting on demand side, the choice literature generally attributes heterogeneity in multi-attribute decision making to different tastes among people, i.e., choice models assume that people have dissimilar levels of preference for the various attributes or attribute levels involved in the choice but follow a similar behavioral rule, usually random utility maximization, to select one option. However, the literature in decision-making reveals that choice is a complex process driven by individuals’ goals and the necessity of balancing the availability of cognitive resources and the requirements of the task. Thus, prior to the resulting observed choice, unobserved latent decisions have been made. These decisions may involve balancing of multiple goals, decisions strategies and choice set formation among other possibilities.

The result is that different individuals follow different choice processes that necessarily shape the multi-attribute choice, i.e., process heterogeneity precedes taste heterogeneity. Choice models that do not account for process heterogeneity may improperly capture such heterogeneity at the taste level.

The demand side will be represented in a utility function described by firm-specific constants, generic attributes and firm specific prices, mimicking a competitive monopolistic market that describes many fast moving consumer goods categories. Taking one of the firms as the focal point for the analysis, an experimental design will allow studying the effects of misattributing process heterogeneity into taste as a function of the relative size of the focal firm’s portfolio, the relative price elasticity of the focal firm and the extent of choice set formation and how it distributes across firms. Choice set formation will be used as the expression of process heterogeneity since it is an outcome of other instances of process heterogeneity in consumer behavior (like goals, decision strategies or decision rules). Thus, demand will arise through a mixture of choice set formation rules. Furthermore, the true data generation process will have taste homogeneity, not heterogeneity. A mixed logit model, allowing for taste heterogeneity, will be used to describe the consumers’ preferences and to identify the market structure. This model, as most of the models described in the literature, will account for taste heterogeneity but not process heterogeneity (here, choice set formation).

On the supply side, the firm will deploy efforts to understand the market structure as a mean of making efficient decisions about portfolio and competitive strategies that will match the perceived market structure. We assume the firm is rational, implying that decisions on production follow the proper profit maximization (or cost minimization) rule, given its understanding of market structure and the constraints of the industry structure.

These decisions are expressed in a generic translog production function describing: (a) presence or absence of returns to scale; (b) presence or absence of economies of scale; and (c) transformation and substitution elasticities. We propose that the configuration of the production function may regulate the severity of the consequences of misrepresenting the market structure, i.e. there will conditions described by the production function that lead to more or less dramatic effects on the firm’s financial performance.

The paper will be structured as follows: (i) an introductory section with justification and objectives; (ii) a literature review of the demand side (choice process and taste heterogeneity) and of the firm side (production functions and market strategy); (iii) the empirical research, using a Monte Carlo simulation to build the true and the biased demand representations; (iv) the analysis of the effects of the different demand representation on the firm’s outcomes, through the production function; (v) the final considerations.

This project studies the effects on the firm’s performance of misattributing process heterogeneity to tastes as a function of the properties of the firm’s production function, of its monopoly power and of the extent of choice process heterogeneity on the demand side.

In marketing, the identification of consumer response heterogeneity is the foundation for segmentation (where to compete) and positioning (how to compete). This implies that the strategic decisions at the marketing level are outcomes of the policy makers’ understanding about consumers’ preferences that would shape market structure and competition patterns. The misunderstanding of consumers’ preferences and demand structure make firms’ performance vulnerable as the statistics on new product failure clearly demonstrate. Accordingly to Nielsen’s Breakthrough Innovation Report Europe 2014, 76% of new product launches in Western Europe did not achieve 52 weeks of sales and almost 50% did not survive the 26th week.

The basic proposition of this paper is that misattributing process heterogeneity to taste is likely to lead firms to be driven away from profit maximization through: (i) first, a wrong understanding of demand will be expressed via the biased location of the utility parameters, leading to the inference of wrong policy measures. More important, the misattribution is likely to suggest variance in the taste parameters supporting the existence of segments that are not actually present in the marketplace; (ii) second, the firm will make long term commitments based on its understanding of the market structure. This will result in an updated production (and cost) function that supposedly maximizes profit given the industry structure; and (iii) finally, when the actual demand does not confirm the supposed market heterogeneity, the firm will be forced to make short term adjustments of production (and costs), away from the optimal.

To study the effect of misattributing process heterogeneity to preferences on the firm’s strategic decisions, the project unfolds in two steps. In the first step, a demand representation is built from a Monte Carlo simulation based on knowledge of the true data generation process, absent taste heterogeneity. As a result, a true and a misspecified representation of consumers’ preferences and market structure will be plugged into a production function that shall allow for the estimation of the effects of the misattribution of consumers’ response heterogeneity in the firm’s outcomes. The severity of the consequences will be studied as a function of characteristics that may be observed either on the firm side or on the demand side.

Starting on demand side, the choice literature generally attributes heterogeneity in multi-attribute decision making to different tastes among people, i.e., choice models assume that people have dissimilar levels of preference for the various attributes or attribute levels involved in the choice but follow a similar behavioral rule, usually random utility maximization, to select one option. However, the literature in decision-making reveals that choice is a complex process driven by individuals’ goals and the necessity of balancing the availability of cognitive resources and the requirements of the task. Thus, prior to the resulting observed choice, unobserved latent decisions have been made. These decisions may involve balancing of multiple goals, decisions strategies and choice set formation among other possibilities.

The result is that different individuals follow different choice processes that necessarily shape the multi-attribute choice, i.e., process heterogeneity precedes taste heterogeneity. Choice models that do not account for process heterogeneity may improperly capture such heterogeneity at the taste level.

The demand side will be represented in a utility function described by firm-specific constants, generic attributes and firm specific prices, mimicking a competitive monopolistic market that describes many fast moving consumer goods categories. Taking one of the firms as the focal point for the analysis, an experimental design will allow studying the effects of misattributing process heterogeneity into taste as a function of the relative size of the focal firm’s portfolio, the relative price elasticity of the focal firm and the extent of choice set formation and how it distributes across firms. Choice set formation will be used as the expression of process heterogeneity since it is an outcome of other instances of process heterogeneity in consumer behavior (like goals, decision strategies or decision rules). Thus, demand will arise through a mixture of choice set formation rules. Furthermore, the true data generation process will have taste homogeneity, not heterogeneity. A mixed logit model, allowing for taste heterogeneity, will be used to describe the consumers’ preferences and to identify the market structure. This model, as most of the models described in the literature, will account for taste heterogeneity but not process heterogeneity (here, choice set formation).

On the supply side, the firm will deploy efforts to understand the market structure as a mean of making efficient decisions about portfolio and competitive strategies that will match the perceived market structure. We assume the firm is rational, implying that decisions on production follow the proper profit maximization (or cost minimization) rule, given its understanding of market structure and the constraints of the industry structure.

These decisions are expressed in a generic translog production function describing: (a) presence or absence of returns to scale; (b) presence or absence of economies of scale; and (c) transformation and substitution elasticities. We propose that the configuration of the production function may regulate the severity of the consequences of misrepresenting the market structure, i.e. there will conditions described by the production function that lead to more or less dramatic effects on the firm’s financial performance.

The paper will be structured as follows: (i) an introductory section with justification and objectives; (ii) a literature review of the demand side (choice process and taste heterogeneity) and of the firm side (production functions and market strategy); (iii) the empirical research, using a Monte Carlo simulation to build the true and the biased demand representations; (iv) the analysis of the effects of the different demand representation on the firm’s outcomes, through the production function; (v) the final considerations.


Conference registration is required in order to view papers.