Cluster Analysis is used to classify objects into relatively homogeneous groups called clusters. It enables marketers to target the needs and wants of population segments rather than using a one-size-fits-all approach. Cluster Analysis is also useful in identifying new product opportunities. A company can analyze its product offerings in relation to its competitors and identify potential new products.
The most important step in cluster analysis is selecting the variables to include in the analysis. The inclusion or exclusion of even one or two variables can drastically affect the outcome. For example, if we were analyzing product preferences and building our clusters using demographic variables, the inclusion or exclusion of age as one of the demographic variables could drastically effect the composition of the clusters.
Hierarchical Clustering
Hierarchical clustering is characterized by the development of a hierarchy or tree-like structure. There are two types of hierarchical clustering, agglomerative clustering, which is the grouping of cases into bigger and bigger clusters, and divisive clustering, which consists of dividing or splitting clusters until each case is its own separate cluster. Under agglomerative clustering there are three methods; linkage, centroid, or Ward’s method. These methods represent various ways to calculate the distances between cases. The video below shows a hierarchical cluster analysis using SPSS.
K-Means Clustering
Nonhierarchical clustering is commonly referred to as k-means clustering. In k-means clustering cases are selected based on a pre-specified threshold value from the center and then grouped together (sequential threshold method) or the cluster centers are selected simultaneously (parallel threshold method) or the cases are reassigned to clusters in order to optimize an over criterion (optimizing partitioning method). The video below shows a k-means cluster analysis using SPSS.
Two-Step Cluster Analysis
In two-step cluster analysis we can select either categorical or continuous variables. These variables can then be analyzed and evaluated against additional variables in the output. The video below shows a two-step cluster analysis using SPSS.