Analyze profitability, allocate resources, using CLV

Problem :

• No metric to understand customer profitability.

• Forecast remaining customer lifetime (years)

• Forecast their future value.

• Determine investment and ROI of each customer.

Data Science Challenges :

• Not enough data. We had to create features from very little information.

• Lots of one time, free trial customers posed a challenge. Should we include them and if we do how do we bring them into the study.

Approach/Solution :

1. The first thing was to find the probability of churning and how long the customer will be with the company.

2. We then calculated the lifetime value and future value of each customer.

3. Prescribed the attributes of negative profit customers, so that next time they are avoided in campaigns. We did not recommend to shed those customers, since they had already been acquired.

4. Identified whom to upsell or cross-sell to increase profitability. Offers were sent after this.

5. Allocated optimal investment for different CLV customers. It answered some of the pressing questions, like how much to invest in customers.

A sample graph depicting division of customers into CLV (for explanation purposes)

Results :

• 18% increase in revenue from cross-selling and up-selling medium value customers.

• 11% increase in sales in the top 10% profitable segment simply by focussing more efforts and resources.

• Reduced campaign costs by $12,000.