Reduced customer churn and increased CLTV for an automotive OEM

A Fortune 50 automobiles OEM wanted to realize the enormous revenue potential underlying its customers.

The objective was to develop a pan-India DataMart of customers and leverage advanced analytics to deliver use cases such as:

• Customer segmentation
• Customer churn prediction
• Car service prediction
• Integrated dashboards
• Customer lifetime value (CLTV)
• Personalized retention and acquisition campaigns

TransOrg followed a phased approach to deliver the solution:

Phase 1: Creation of a customer 360 DataMart

Before the analytics intervention by TransOrg, client’s data had many inconsistencies, inaccuracies and duplicate entries rendering it unusable for data analytics.

TransOrg centralized, cleaned, and de-duplicated the customer data to build the customer 360 DataMart which gets updated after every 24 hours. As part of this exercise, TransOrg generated a single CustRelID (relationship ID) for each customer by stitching together data from various sources such as:

• Sales data
• Service data
• Used car sales data
• Customer relationship data
• Enquiry data
• Sales app (transactional and interactive data)
• Customer connect app
• Vehicle configuration virtual platform app

After cleansing and stitching, DataMart consisted of over 0.2 million customers, 2.5 million transactions, and 0.1 million vehicles.

Phase 2: Advanced analytics on DataMart

TransOrg leveraged the customer DataMart for various advanced analytics applications:

1. Database cleansing and customer de-duplication

TransOrg developed sixteen distinct de-duplication rules and built a customer de-duplication engine that in real-time identifies an existing customer by assigning a unique customer identifier code called the CustRelID to each record.

2. Integrated dashboards

Created integrated dashboards as a unified source of information from every touch point such as enquiry, booking, delivery, service, etc. with various KPIs pertinent to business users across the organization hierarchy from the sales executive at a showroom to senior management.

3. Customer segmentation

TransOrg used K-means clustering technique to derive customer segments based on customers’ demographics, purchase behavior and loyalty points to create segment specific campaigns enhancing customer delight and boosting after-sales revenues.

4. Customer churn prediction

TransOrg adapted ensemble models to predict customer churn probability rate for different customer segments. TransOrg identified drivers behind customer churn such as:

• Vehicle model
• Lifetime service attributes
• Post-service feedback

5. Customer lifetime value modeling

TransOrg calculated the customer lifetime value (CLTV) for each customer using the following parameters:

• Churn Probability Rate or the probability of a customer churning out of the aftersales service network
• Customer Value or the value of a customer’s purchases over a timeframe e.g., over the last five years
• Customer Lifespan or the length of time a customer continues to buy new or replacement cars from the client
• Discount Rate or the rate of return used to discount future cash flows back to their present value

• Reduced duplicate customer profiles by 45%
• Increased response rates of targeted campaigns for multiple products by up to 48%
• Increased revenues by 5% by targeting high CLTV customers with upselling and cross-selling campaigns
• Improved decision making with insights from integrated dashboards