Typical Problem Studied with Cluster Analysis: Probably the most typical application of cluster analysis is found in market segmentation studies. Many firms try using different attitudes to identify the different segment in a given market. Such situations are quite different from the ones where researchers try to identify different market segments on the basis of one variable only (e.g. heavy users/average users / light users, or prefer Brand A/ prefer Brand B). When researchers are trying to segment on the basis of several attitudes or variables, they are likely to use cluster analysis rather than one of the method.
An example of such segmentation is a sporting goods company that wishes to identify the various market segments that constitute the total market for sporting equipment. A large sample of users of all kinds of sporting equipment are asked to report their attitudes regarding their preference for indoor or outdoor sports., their preference for rugged or easy sporting activities, which sports they prefer, and so on. Cluster analysis can be used on these data to see if the total market consists of a number of different segments.
What Cluster Analysis Does: Cluster analysis identifies different groups (or clusters) of respondents, such that the respondents in any one cluster are similar to each other but different from the respondents in the other clusters.
Cluster analysis is typically applied to data consisting of many variables that have been collected from a large sample of respondents. The cluster analysis procedures search through the data and identify respondents who have given identical, or at least identical, or at least similar, answers to a certain combination of questions. These respondents are formed into one cluster.
The cluster analysis procedures then search through the data looking for a second set of respondents, all of whom have given similar answers to some other combination of questions. This second set of respondents is all similar to each other, but they are quite different from the respondents in the first cluster. By proceeding in this manner, the cluster analysis procedures may identify a third cluster of respondents who are different from the first two clusters. The procedures can be continued until of the different clusters have been identified.
Types of Variables Used in Cluster Analysis:
Cluster analysis is typically applied to data that have been recorded on interval scales such as 5 –, 7 –, or 10 – point scales, but it can also be applied to continuous variables data. However, there tends to be fewer applications of cluster analysis to continuous variable data than to scaled data.
Cluster Analysis Identifies Interdependencies among Variables: Each of the previously discussed multivariate methods (cross tabulation, regression, LDA and AID) was concerned with a single variable that in some way was important to marketing decision makers. For example, in the LDA illustration the single variable of interest was the salesman’s success, and in the AID illustration the single variable of interest was the number of advertising and other information sources used when purchasing appliances. In those cases the researcher was using a single variable to identify a class or a category into which a respondent belonged.
There are situations in which a researcher may wish to use more than one variable to identify the class or category into which a respondent belongs. A multiple variable classification is used whenever it is more useful to marketing decision makers than a single variable classification. When researchers encounter such situations, they will want to use cluster analysis or one of the other two methods yet to be discussed. This is indicated in the bottom half, which shows that cluster analysis is concerned with finding interdependencies among a number of variables that were measured in the study especially within different subsets of respondents. Readers should note this because it will help them better understand when cluster analysis can be used rather than some other method.