ASYMMETRIC CLUSTER ANALYSIS OF BRAND SWITCHING AMONG SOFT DRINKS

摘要

Brand switching data among eight soft drink brands were analyzed by using an asymmetric cluster analysis. The eight soft drink brands consist of four diet and four non-diet brands, and of four cola and four non-cola brands as well. The brand switching data matrix was rescaled to ease the effect caused by the differences in the market share of the brands. The elements of the rescaled brand switching matrix were regarded to be proximities among the brands, and the matrix was analyzed by an asymmetric cluster analysis. The asymmetric cluster analysis used in the present study is able to represent not only asymmetric relationships of the brand switching but also self-clustering or clusters of stayers of a brand on the resulting dendrogram. The obtained result showed that the royalty toward non-diet soft drink brands are stronger than that toward diet brands and that the differences between the diet and the non-diet brands were larger than those between cola and non-cola brands.