In recent years increasing attention has been paid to segmentation based on the benefits wanted from a product. This is particularly true with new products when the issue is often whether to develop a separate product for each benefit segment. Clustering is based on the wants prospective consumers have regarding the characteristics of various new product concepts when compared with existing brands. Essentially, this procedure for predicting purchase behavior is based on a theory that holds that an attitude toward a brand is made up of beliefs about the brand’s attributes combined with the importance given those attributes. Thus, “a consumer’s attitude toward a particular brand in a certain product is hypothesized to be a function of the relative importance of each of the relevant product attributes and the consumer’s beliefs about the brand on each attribute. A score is obtained by scaling the extent to which each brand is perceived as possessing certain characteristics and the importance attached to each by the consumer. Based on the scores for each brands for each brand, a prediction is made of the individual’s preference ranking.
In its complete form such a segmentation model will show the consumer’s ideal combination of product characteristics (each of which is weighted as to its relative importance). Clustering of those respondents with similar “ideals” enables the researcher to belief to better understand not only what consumer want, but how the various brands and new product concepts are perceived with regard to the ideal brand.
To illustrate, assume that respondents in a sample are asked to describe their ideal make of car by rating the importance of a number of different attributes that collectively serve to explain differences between brands. Each make is rated using the same set of attributes. Given such data and using multiple discriminant analysis, it is possible to determine the weighted combinations of attributes that best distinguish respondents’ ideal brands from actual brands.
The above procedure can be illustrated in a figure where consumers are grouped into five segments based on their similarity with respect to the ideal brand. The size of the circle represents the relative proportional of consumers within a particular cluster. Thus, the first segment is the largest, and the fifth segment is the smallest. Also shown is how consumers position the various makes or brands of cars.
By comparing the preferences of different segments with their positioning of the various brands, much can be learned about the competitive strengths of the different brands, the extent to which the brands are in competition with one another and the opportunities for new products. For example, note that BMW has a strong and relatively unchallenged competitive position among customers in the fifth segment while Ford, Chevrolet, and Toyota seem to be in intense competition with each other for the preferences of the customers in the first segment. None of the brands seems to be strong relative to the preference of customers in the third segment and thus there may be an opportunity for a new brand.
Such an analysis as indicated above should prove helpful to a company in determining what products to support and to what extent. The share of a particular product should be increased by repositioning closer to the target segment and further from other brands.
Conjoint Analysis: This method helps management better understand the ways in which product features are linked to product preferences. It does so in the form of utilizes provided in numerical form for different levels of the product features involved. For example, in producing a new truck, designed to achieve a given position, we would need to know what mix of physical characteristics, including their level (e.g. size of cab, length of body, maximum load, and miles per gallon) are most desired. We would also needs to know how to make certain tradeoffs as, for example between overall weight, size of cab and miles per gallon.
Conjoint analysis is particularly useful in new product development work since it literally determines which combination of attributes is most preferred out of all possible combinations. Thus, several new products (based on different combination of attributes) can be tested to determine which is preferred by the largest number of respondents and to identify the characteristics and to identify the characteristics of that market segment preferring a particular product.
In recent years conjoint analysis has become increasingly popular as management attempts to better link product attributes with the needs of certain market segments. A number of large companies such as General Motors, Ford, General Electric, Xerox and general Foods have used this approach across a wide range of products.