Case · Telecom
Mobile carrier — churn prediction & plan-impact simulation
// PERSONA
Lim, Subscriber Analytics — mobile carrier
// INDUSTRY
Telecom
// DATASET
case_10 · subscriber_usage_log.csv · 4.2M subscribers
PROBLEM
The Problem
With 4.2M subscribers, churn signals had to be caught proactively; without per-variable attribution, retention actions scattered and wasted budget. And before launching a new plan, the impact on revenue and churn had to be simulated — but no such tool existed.
COMING SOON
APPROACH
XimTier Approach
Subscriber usage, service quality, and plan history are analyzed end-to-end, with SHAP-based per-variable contributions auto-attached to every churn score. Reverse What-If pre-simulates a new plan's revenue and churn impact so launch decisions ship with auto-generated justification.
COMING SOON
OUTCOMES
Outcomes
OUT / 01
Proactive care of high-risk subscribers (churn −16%)
OUT / 02
Auto-identified drivers of network quality drops
OUT / 03
Pre-simulated impact of new plans
OUT / 04
Retention-campaign ROI +43%
⚠ Example data — fine-tuned on customer data at deployment.