Recently, I was meeting with a DB2 for z/OS client, and the topic of Next Best Action (NBA) came up.
My client's challenges are that although they consider the "lifetime value" of the customer in their marketing messaging and fraud detection algorithms:
(1) Marketing messages are both segment based (or shotgun blast) and often ill-timed for the customer based on their value and behavioral lifecycle. Customers receive multiple outbound messages per month.
(2) In many cases, automated interactions with the customer for upsell, cross-sell, and fraud detection are based on scoring input data that is aged one or more days, not representing the current state of the customer
(3) Customer interactions are initiated via separate organizations with limited to no coordination nor cross-channel understanding of the customer.
Unfortunately, the result of these poorly targeted and ill-timed interactions with customers is that the customers feel like they are treated with little consideration of their history with my client. One wrong or ill-timed interaction is all it takes to destroy many years of relationship building with a customer and send their lifetime value into a death spiral.
NBA is all about taking the right action with the specific customer via the right channel at the right time based on a cross-channel view of their behaviors and value. In short, it is mass automation of the one-to-one relationship with the customer. It leverages a combination of automated rules-based decision making, mathematical optimization models, in-transaction and batch scoring, as well as integrated campaign management.
For more details on NBA, here are a several resources to get you started: