Improve customer retention with churn analysis and machine learning

Customer attrition, or churn, is one of the most important challenges faced by modern enterprises. The problem aspects a growing number of industries. As the competition grows, so do customer expectations.

Wanting better product quality at a lower price is just one example of a case where a customer may choose one company’s offering over that of another. Gaining insights into customer satisfaction and risk factors helps to align strategy with customer expectations and to reduce customer churn.

Challenge

The company is a multinational commercial car manufacturer. Their customer demographic is multifaceted. In order to efficiently direct their marketing, they needed deeper insight into their consumer base.

We recommended a customer churn analysis solution to identify customers who are less likely to switch to a competitor. This way, the client can focus their strategy on where it’s required, increasing customer retention while saving money.

To increase customer retention, we needed to implement a solution that provides the following insights:

• Analysis and monitoring of customer satisfaction and loyalty.
• Monitoring the number of customer departures and detecting trends.
• Examining customer behavior patterns, like what actions precede customer attrition.
• Defining customer segments and determining which are at the highest risk of leaving.
• Using machine learning models to calculate the probability of a particular customer leaving.
• Analyzing customer revenue and taking into account various dimensions like time, geography, product models, etc.
• Researching and understanding customer purchasing patterns.

Solution

We created a customized analytics system using technologies like Microsoft R Open, Apache Hadoop, and Tableau.

Our consultants first analyzed the data for its application and validity to forecast customer migration. Then we built machine learning and Tableau models. The models were tested and selected for the one most suitable to define customer expectations.

We implemented clustering algorithms to segment customers based on various parameters. Finally, we visualized the resulting information in Tableau.

Results

Data visualizations enable the client to see trends in a friendly format.  Now, end-users can effortlessly view reports and quickly scan them for signs of customer attrition. The reports also include features like drill-down analysis and trend forecasting. This helps our client to identify segments to focus their marketing strategy on.

New insights enable the client to provide higher quality and more relevant services, and thus, save on marketing costs, increase customer retention, and reduce churn.