Using Predictive Analytics to Reduce Churn and Boost Sales

Trident Marketing acts as a dealership for DirecTV, whose products represent a USD40 million piece of Trident’s business. However, if a DirecTV customer cancels within six months of signing up for the service through Trident, the company must refund DirecTV all of its commission fees—which makes customer churn an expensive proposition.

Average customer churn in the industry is typically 2.5 to 3 percent. As Trident’s customer base grew, however, its churn rate surpassed 4 percent and continued to grow. As a result, DirecTV challenged the organization to cut churn in half or risk losing its dealership agreement.

Trident Marketing CIO Brandon Brown turned to IBM Premier Business Partner Fuzzy Logix to help his company analyze and model prior sales. By using its own in-database analytics software with an IBM® Netezza® Data Warehouse appliance, Fuzzy Logix enabled Trident to model the sales decision process overnight.

Today, customer churn is half of what it once was. Trident can now make real-time decisions about which customers to approve for new accounts and how much each customer should be charged based on his or her risk level. The company can even make real-time predictions about which products a customer is most likely to purchase next, which has helped to increase upsell rates and enabled the company to achieve a return on its investment within 12 months—much faster than the 36 months originally projected.

To learn more, read the full case study here.