Customer Segmentation Automotive

Challenge

Analyzing and modeling customer groups to identify new customer needs is a central task in automotive market research. As part of annual panel surveys, thousands of individuals are asked about their desires and needs regarding automobility. These data are systematically analyzed by automotive OEMs to identify new customer needs in a timely manner. Cluster analyses are used to uncover customer segments that possess homogeneous characteristics within themselves but differ as much as possible from other groups. Our client, an international automotive corporation, faced the challenge that previous cluster analyses produced highly volatile results that varied significantly from year to year.

Approach

Based on the panel survey data provided, we first replicated the client's approach as a benchmark. This revealed unsatisfactory cluster separation. The reason for this was the extreme heterogeneity among the constructs used. After data cleaning and dimensionality reduction of the various constructs, separate cluster analyses were conducted within each construct. The results of the subclusters were then processed in a meta-cluster analysis, which showed good to very good discriminants of the cluster solutions. The substantive interpretation of the results also yielded plausible outcomes.

Result

With the newly developed two-stage clustering method, our client can now identify more stable customer clusters that exhibit significantly stronger separation than the previous solution. The newly formed clusters help our client identify new customer groups and their needs, testing them within the framework of new product strategies.