In this project, we distinguished and found similarities between different customer groups using Cluster Analysis and Discriminant Analysis for a hypothetical company based on a credit card dataset.
Major corporations and large size retail outlets often have difficulty keeping track of the buying habits of every customer, but it is important that businesses
keep a consistent base of customers and simultaneously try to expand the customer base, but most major corporations and large size retail outlets often have difficulty keeping
a track about the buying habits of every customer. To achieve that, we studied the ‘best’ customers i.e., the regular, high-spending customers as well
as the high-risk of churning or ‘dormant’ customers. This prevents the company from spending its limited resources on customers that are likely to churn and prevents
valuable customers from churning by offering them personalized marketing campaigns.
We used the customer credit card behavior to understand their transactional patterns and hence, target them with campaigns accordingly. We used the K-means clustering algorithm to identify the customer segments from the variables given in the dataset and Linear Discriminant Analysis to make a comparison of results from K-means clustering and see if the variables chosen help in distinguishing different customer segments.
With these results, we made marketing recommendations based on the customer segments achieved. In detail, we will discuss, what kind of customers we are encountering and how the retail company can market to them.