CHURN DETECTION & PREVENTION
M is a multinational corporation, which develops, manufactures, licenses, supports and sells computer software, consumer electronics and personal computers and services. It is one of the world's largest software providers, and also one of the world's most valuable companies.
The customer satisfaction survey results are in decline. M has a wealth of survey data but not sufficient insight into the drivers of satisfaction and churning behaviour. A number of hypotheses exist as to the decline. Additionally, an ideal customer experience helps to retain customers. Indicators of significant decline are driving the requirement to aggregate all the data available to build a predictive analysis model and strategy recommendation engine. Taking a predictive approach allows M to take action in advance of a potential defection and customer churn. It is particularly important for M to identify those customers who are dissatisfied and wish to transfer their habit of buying products and/or services to compete other merchants. Accurately predicting customer behaviour is a challenging task. On the other hand, finding the best strategies to prevent customer defection is even tougher.
Most customers warn you before they leave. Therefore, the core task for churn detection/satisfaction ranking is to capture the "warning" from the big data. Business data from a broad range of domain are incorporated, such as call center, sales, revenue, maintenance, social media, email text, and online system web log. For the first time within M, big data analytics techniques have been leveraged to prepare, combine and analyse the merged data set to uncover insights hidden in the data, and build the predictive models and recommendation engines. The approach employed in the behavior analytics including (1) contrast analysis which discovers the difference between satisfactory customer experience/interaction and unsatisfactory one; (2) behavior analytics, which allows M to find out the driving force of churn; (3) causal rule mining, automatically generates the action plan for individual case that may lead to a churn; (4) deep learning, to predict the trend of churn for a given customer group; (5) big data visualization, to present the decision maker the overview of churn risk and interesting behavior trend.
Deliverables & Benefits
- Reduced Churn Rate and Increased Revenue With the proper early intervention, those unsatisfactory customers can get timely help and special care. All churn "warnings" are visualized and monitored, once the risk threshold is reached, interaction alerts will be triggered. Furthermore, optimal intervention strategies are also proposed by the recommendation engine.
- Satisfactory customer experience and good competitiveness We are able to dig deeply into every aspect of the process that consumers experience, and provide M with detailed recommendations for satisfaction improvement and churn prevention. We deliver these insights through big data analytics and visualisation.
- Reduced customer maintainence cost Costly manual process was replaced by the maintainence engines driven by intervention strategies that are automatically generated and executed, such as email notification, eGift card, special offers, etc.
- Visible risk control Rather than dealing with a black box of risk indication, the potential risk of customer churn is monitored closely and presented to the decision maker in a business-friendly visual interface.
- Cost-effective business operation Both long term and short term dollar values are fully considered in assessing the improvement of detention rate and satisfaction level of the customer.