How to predict customer churn in every interaction

Manager of Portfolio Strategy, IBM

Customer satisfaction is as old as commerce itself—indeed, keeping customers happy and engaged helps keep the economy going. But research shows that in industry after industry, customer acquisition comes at significant cost. What’s more, years may pass before a customer becomes profitable after having been acquired. Accordingly, increasing customer retention rates by 5 percent can boost profits by as much as 25 percent to 95 percent.

Remember the old adage: A satisfied customer tells nine other people about a positive experience, but a dissatisfied customer may talk to as many as 22 other people about a negative experience. Are you heavily invested in client relationship management (CRM) solutions? Have you spent years consulting with experts to build solid propensity to buy models and other analytics? Certainly such systems are vital to a business, and they serve as the backbone of operations—but how much data do these systems capture out of the universe of data available? What’s more, is your CRM solution versatile and responsive? Or perhaps you feel as if you are stuck in the 1990s—and how do you work with voice data or images?

Certainly it’s no secret that you can use data to help you predict customer churn, but that’s only a beginning. How can you go on to act disruptively, engaging with your data and your clients in unexpected ways? In short, why isn’t everyone disrupting on the scale of Uber, AirBnb or Alibaba?

To gain insight into the answers to these and other questions, join us for the Predicting Customer Churn in Real Time” webcast, during which we will look at how to enhance our predictions of customer churn by thinking about how we use data. We will also examine how you can boost the usability of your data, as well as ways of helping you and your clients engage with your data. 

Get started with these insights

  • Begin by looking at technologies that extend the warehouse: NoSQL databases, graph databases and Hadoop. Use these approaches to gather the data available to you, such as social data or images, without taking on the burden of building data models—for starters, you can use Hadoop to store both structured and unstructured data.
  • Adopt a new data science practice: Tell stories using all available data. With the aid of a modern repository, data science professionals can explore data in new and exciting ways. Let building skills in R, machine learning and Spark form a palette of sophisticated analytics that goes beyond Boolean logic and if-then-else constructions.
  • Engage clients on their data: Disruptors such as Uber and AirBnB not only bring data to marketing teams to help them create offers and campaigns, but they also offer ways for their clients to engage with data. Imagine, for example, your airline of choice consulting you about how best to reroute your flights in case of delay—or giving you miles for every extra minute you spend on the runway waiting to take off. Indeed, preventing churn isn’t just about pushing offers; it’s about engagement—about empowering your clients to use and work with data.
  • Respond to events as they unfold: Captured events or data streams, along with streaming analytics, can give you up-to-the-second insight into clients, just as voice analytics can help call center reps serve disgruntled callers. Similarly, analyzing data streams from mobile devices can not only help you customize offers to recipients’ geographic locations, but also enable you to respond to unexpected usage events.

We are thankful for our clients, but are they thankful for us? Join us for this webcast to learn how you can give your clients reasons to be thankful for you and your products and services all year long. Also, to learn more about Hadoop, visit