Research Tips

Multidimensional Scaling

Understanding how customers think about products or services within a particular category is important to manufacturers, retailers, and service providers, but it’s often difficult to determine what particular combination of attributes buyers use to evaluate complex products and services. For example, women’s apparel could be classified according to color, fabric, designer, size, brand, brand image, type of garment, normal use of garment, etc. Someone trying to understand competition within such a product category could make an arbitrary selection of one or several attributes and group competitors on that basis, but multidimensional scaling offers a better alternative. Multidimensional scaling is an analytical technique that enables inferences to be made about which attributes buyers use to distinguish competitors from one another. It does this by starting with similarity ratings (buyer ratings of how similar one brand is to another — the buyers don’t need to say why they think one brand is similar to another, only that they do) and using statistics as the basis for inferences about the underlying dimensions that explain that similarity.