ML Powered Proactive Collection Program

Prior Situation / Scenario:

• Churn scores were ineffective since they predicted the churn event too late
• Retention campaigns were reactive
• Retention offers were costly and same for all segments
• Reactive churn campaigns

Client Challenges:

• Improve involuntary churn rates. Reduce bad debt
• Retention & collection process was manual and not segmented
• Costly and ineffective offers to retain customers
• Lack of harmonized retention offers
• Contact Center retention calls with substantial costs

Strata Solution/ Key Enablers:

• Developed two “Propensity to Involuntary Churn” risk models and a new customer value segmentation (Over 10M customer records and more than 550 variables analyzed)
• Defined new collections process and actions that combine value, risk and created omnichannel contact journeys (Call Center, IVR, Email, SMS) to optimize costs and results based on value, risk and prior contacts
• Early detection of churn predictor events (wait time, claims, service, visits)
• Automated retention processes and retention matrix

Outcome:

• Identified a sequence of events that were incorporated as churn predictors in a propensity to churn algorithm that activates retention treatments and actions
• Design and implementation of pyramid approach segmentation criteria for stimulation campaigns optimizing contact/conversion/impact
• Turning retention campaigns into proactive actions
• Unified retention treatments by customer value segments
• Bad debt reduced with better identification of customers