Customer Churn & Retention Intelligence
Predictive subscriber attrition modeling and interactive Power BI executive dashboard
the problem
A subscription SaaS service was experiencing an annualized churn rate exceeding 24%, with customer success teams unable to pinpoint why customers cancelled until exit surveys were submitted. The leadership team lacked visibility into leading indicators of churn and needed an automated early-warning framework.
approach & architecture
Engineered an automated data extraction and cleaning pipeline from PostgreSQL transactional and behavioral logs. Conducted in-depth exploratory data analysis (EDA) across contract types, payment channels, and monthly usage metrics. Trained and validated classification algorithms (Random Forest & XGBoost) to generate churn probability scores, and built a dynamic Power BI report with customer risk tiers for intervention.
system highlights
key features
Early-Warning Risk Scoring
Segments the active customer base into High, Medium, and Low risk tiers based on real-time activity and contract age.
Driver Attribution Analysis
Feature importance analysis revealed month-to-month contracts and electronic check billing had 3.2x higher churn rates than annual plans.
Interactive What-If Scenario Modeling
Power BI parameter sliders allowing finance and marketing leaders to model revenue preservation based on incentive discounts.
Automated Data Hygiene & ETL
Standardized Python scripts that validate incoming data schemas and output pre-calculated metrics for downstream reporting.
technologies
outcomes & metrics
Transformed reactive customer cancellation into a proactive retention engine, equipping decision-makers with quantified customer health scores and actionable intervention playbooks.
18% reduction in customer churn within 6 months of campaign launch
$280K+ in annualized recurring revenue saved through proactive renewals
86% recall rate on identifying at-risk accounts 30 days prior to contract renewal