Some difficulties in predicting new product trial using concept test scores

Abstract

At an early stage in the new product development process, marketers often evaluate several concept statements in terms of customer preferences to choose the best concept for further development. Purchase intention scale is often used to measure consumer preferences at this stage when the product is still a concept statement or a mathematical position on a perceptual map. This paper discusses the limitations of two methods of aggregating individual preferences, namely plurality (first-choice) and the Condorcet (pair-wise majority) methods. The plurality method is subject to the top-box paradox while the Condorcet method suffers from the paradox of voting. The Copeland method is presented as an alternative to the Condorcet method when the latter fails to identify the majority's choice. Some limitations of predicting product trial are also presented.Predicting new product trial for a choice set Consumer packaged goods marketers utilize mail interviews or mall/in-store intercept to measure respondents' intention to try the product. The most Evaluates three rules

Are Student Groups Dysfunctional?

“…The top-box score is the percentage of respondents who marked the highest rating on a scale. This technique is commonly used in marketing research (e.g., Bohling et al, 2006;Dubas, Dubas, & Atwong, 1999). For example, on a 7-point scale, the topbox score would be the percentage of respondents who marked a 7 on the scale.…”

An Illustration of Potential Sources of Concept‐Test Error*

“…Because the determination of whether a concept should go forward or not is firm specific, results of each concept are interpreted relative to others. See, for instance, Dubas, Dubas, and Atwong (1999) and Trebbi and Flesch (1993) for a similar approach.…”

“…As mentioned, concept testing can provide an initial estimate of a candidate product's sales (Crawford and Di Benedetto, 2003). This estimate is typically derived from the key measure of purchase intent (Dubas, Dubas, and Atwong, 1999) and also from measures of liking (Cooper, 1993;Iuso, 1975). Often, at least with academic studies, liking and purchase intent scales are combined to arrive at an aggregate measure of customer response (see, e.g., Keller and Aaker, 1992;Klink and Smith, 2001;Moreau, Lehmann, and Markman, 2001).…”

“…To illustrate how these concerns can affect the interpretation of concept‐test results, the present study replicates in many ways a conventional concept test. It gauges the market response to several product concepts and thereby produces results similar to what a decision maker, or manager, would use to select one or more alternatives for further development (Dubas, Dubas, and Atwong, 1999; Page and Rosenbaum, 1992).…”

Whose Crystal Ball to Choose? Individual Difference in the Generalizability of Concept Testing

“…Most recently, Klink and Athaide (2006) summarize the three basic design issues inherent to concept testing as stimuli design, respondent selection, and response measurement. While there is considerable academic literature on stimulus‐related issues (e.g., “concept formulation” in Lees and Wright, 2004; “competitive‐set information” in Miller, Bruvold, and Kernan, 1987), and response measurement (Dubas, Dubas, and Atwong, 1999; Jagpal, Jedidi, and Jamil, 2007; Jamieson and Bass, 1989), less academic attention has been given to subject selection in concept testing (Klink and Athaide).…”