One of the main goals is to increase customer LTV, reason why we need to better understand them. Clustering our customer database is one of the first steps to do so.

My solution was to implement an RFM model (recency, frequency and monetary) to segment our customer database into multiple segments. Also implemented a k-means algorithm to compare both clustering outputs.

Having a segmented customer database we could easily spot customers based on their previous behavior and prioritize customers based on their attributes.

This clustering methodology also helped design business goals on how we'd like to move customers from one cluster to another, aiming for customers to increase the number of orders, their frequency and LTV.