A naukri.com initiative
Medium
12h
281
Image Credit: Medium
☎️ TELCO CUSTOMER CHURN PREDICTION
- A dataset with 7,043 records was used for telco customer churn prediction, focusing on various factors like usage patterns, contract types, and service complaints.
- Steps included data cleaning, EDA to uncover churn drivers, preparing data for model building, and constructing a Random Forest model.
- EDA revealed key churn drivers such as monthly charges, tenure, and contract types, with month-to-month contracts showing higher churn rates.
- The Random Forest model achieved around 80% accuracy despite data imbalance, highlighting predictors like tenure, charges, and contract types.
- Proposed strategies to reduce churn include promoting longer contracts, optimizing pricing, enhancing service quality, refining payment processes, and investing in predictive analytics.
- Recommendations involve incentivizing longer contracts, adjusting pricing for high-cost services like fiber optics, improving tech support, simplifying payment processes, and utilizing predictive analytics for real-time insights.
- Overall, the project emphasizes actionable strategies to address churn drivers, enhance customer loyalty, and drive sustainable growth.
Read Full Article
16 Likes
For uninterrupted reading, download the app