Alternative approaches to cluster-based market segmentation
The effectiveness of several clustering techniques commonly used in marketing research was evaluated. The techniques were implemented on nine large-scale data sets by means of a k-means clustering routine which is highly similar to BMPD KM. Results indicate that clustering based on rotated factors led to data which have poorer associations with exogenous variables. On the other hand, clustering methods based on the original data were found to be capable of generating data with that show association with exogenous variables.
Publication Name: Journal of the Market Research Society
Addressing consumer preferences in brand choice is of utmost concern to a supplier. A multidimensional approach of combining perceptual mapping of consideration sets (consumers' choice brands) to numerical testing, where a consumer's closeness to brand is tested, is discussed. Analysis of consideration sets not only helps consumers in their purchasing decisions, but also helps suppliers in marketing the most preferred brands of their customers. Interbrand similarity and its statistical significance is determined.
Publication Name: Marketing Research: A Magazine of Management & Applications