Comparison of PLS dummy variables and Fishbone method to determine optimal product characteristics from ideal profiles

Abstract

Sensory professionals mostly used trained or expert panels for diagnostic purposes and only use consumers for hedonic assessments. In market research, consumers are not only used for hedonic assessments, but also for product diagnostic purposes (as is the case with Just About Right procedure). In the Ideal Profile method, consumers are also used for both tasks. They rate the perceived and ideal intensities and the acceptance of a series of products. The two sets of information (product description and hedonic data) are then used for product improvement. In this paper, the results of two methods of analysis – PLS on dummy variables and Fishbone method – will be compared. The objectives of this study were (1) to compare PLS and Fishbone method to determine their similarity in predicting the impact of the attributes on overall liking, and (2) to determine if the methods would return similar or contrasting conclusions. These methodologies have been applied to a study concerning 12 commercially-available women’s perfumes. Though the differences in model derivations caused some small dissimilarities, similar trends were found between products across methods for those perfumes far from the ideal. There was agreement regarding which attributes are too strong or too weak, and the order of importance of these attributes for liking. The two methods showed greater dissimilarity for perfumes that were already near the consumer’s ideal.