Cluster Analysis 9

Validate and Interpret the Clustering Solution

Before interpreting the cluster solution, we need to assess the stability of the results. Stability means that the cluster membership of individuals does not change, or only changes a little when different clustering methods are used to cluster the objects. Thus, when different methods produce similar results, we claim stability.

The aim of any cluster analysis is to differentiate well between the objects. The identified clusters should therefore differ substantially from each other and the members of different clusters should respond differently to different marketing-mix elements and programs.
Lastly, we need to profile the cluster solution by using observable variables. Profiling ensures that we can easily assign new objects to clusters based on observable traits. For example, we could identify clusters based on loyalty to a product, but in order to use these different clusters, their membership should be identifiable according to tangible variables, such as income, location, or family size, in order to be actionable.
The key to successful segmentation is to critically revisit the results of different cluster analysis set-ups (e.g., by using different algorithms on the same data) in terms of managerial relevance. The following criteria help identify a clustering solution (Kotler and Keller 2015; Tonks 2009).

• 
  Substantial: The clusters are large and sufficiently profitable to serve.
  
• 
  Reliable: Only clusters that are stable over time can provide the necessary basis for a successful marketing strategy. If clusters change their composition quickly, or their members’ behavior, targeting strategies are not likely to succeed. Therefore, a certain degree of stability is necessary to ensure that marketing strategies can be implemented and produce adequate results. Reliability can be evaluated by critically revisiting and replicating the clustering results at a later date.
  
• 
  Accessible: The clusters can be effectively reached and served.
  

⁷See Punj and Stewart (1983) for additional information on this sequential approach.

• 
  Actionable: Effective programs can be formulated to attract and serve the clusters.
  
• 
  Parsimonious: To be managerially meaningful, only a small set of substantial clusters should be identified.
  
• 
  Familiar: To ensure management acceptance, the cluster composition should be easy to relate to.
  
• 
  Relevant: Clusters should be relevant in respect of the company’s competencies and objectives.
  

1.