Case Study: UPC Switzerland achieves 34% reduction in customer churn with GemSeek

Reducing mobile operator customers churn by 34% with predictive analytics

UPC Switzerland, part of Liberty Global, had a successful customer rescue program but could only reach about 15% of dissatisfied customers via surveys, limiting scalability. GemSeek was engaged to extend that program by using customer analytics and predictive analytics to identify unhappy customers among the remaining 85% who never responded to surveys.

GemSeek built a gradient-boosting machine learning model that mapped survey respondents to CRM and usage data (1,200 variables) and scored the full customer base, then embedded the model into UPC’s operations so customer care could proactively target at-risk subscribers. The approach cut churn in the treated population by 34% (8% vs. 12% in control), delivered a 500% ROI in year one, identified super-detractors with high precision (92%), raised NPS by 3 LTR points, and transformed the call centre into a revenue-generating unit.

UPC Switzerland