The dirt on bikes: an illustration of CART models for brand differentiation

Abstract

Purpose – The primary purpose of this study is to illustrate a statistical method to identify product attributes that differentiate branded products from those of competitors. Design/methodology/approach – The authors use classification and regression tree (CART) models in an analysis of observable characteristics of a mature category of relatively complex products, dirt bikes. Findings – The authors show how the CART model can be used as a tool for identifying brand differences and to summarize product categories in terms of these key differences. Research limitations/implications – The work focuses on physical specifications of the products at one point in time. An important area for future extensions will be to incorporate consumer utility into the analysis. Practical implications – The approach will offer value to brand managers and product managers who have a goal of maintaining the alignment of the brand with the underlying observable differentiation of the branded products. The approach can also serve as the basis for a product/brand performance report (similar to consumer reports) by identifying a select set of product characteristics that differ across brands. Originality/value – Products serve as influential sources of information about a brand's identity. To the extent that observable product characteristics do not match brand claims, consumers may question the brand's authenticity. Although for some products it may be a reasonably simple task to identify the set of observable product attributes that have implications for the brand identity, the task can be challenging for many products. The authors employ an analysis technique to reveal product characteristics that are consistent within brand product lines but that differ across brands.