Retention Intelligence

Real-time churn monitoring and predictive analytics

🌙
AD

🤖 ML Prediction Engine — Ready for Model Integration

Random Forest model integrated · 99.5% accuracy · Run python app.py to activate live predictions

↑ 2.4%
Overall Churn Risk
53.7%
537 of 1,000 customers
Total: 1,000
At-Risk Customers
511
Churn probability > 50%
↑ High
Potential Revenue Loss
₹2.71 Cr
From 537 churned customers
↑ 12%
Retention Success Rate
46.3%
463 customers retained
Churn Risk Trends
Monthly churn rate across the dataset
Churn by Segment
Membership & region breakdown
Gold Members52.7%
Silver Members54.2%
Bronze Members54.2%
East Region57.8%
West Region48.9%
Top At-Risk Customers
Highest churn probability — immediate retention action required
Customer Name Risk Score Churn Probability Satisfaction Days Inactive Recommended Action
All Customers
Showing all 1,000 customers
Customer Churn Prob ↓ Risk Score Membership Region Satisfaction Days Inactive Status Action
1–15 of 100
🎁
Personalised Discounts
Auto-trigger vouchers for at-risk customers. Threshold-based targeting via churn probability score.
Discount Amount
20%
Targets511 customers
Est. Recovery₹45L+
📞
Follow-up Calls
Assign personal outreach calls to account managers. Priority-based assignment by region and risk.
Priority Threshold
Agents Available12 agents
Success Rate68%
⭐
Loyalty Rewards Boost
Double loyalty points for at-risk customers. Proven to raise satisfaction scores by an average of 1.2 pts.
Campaign Duration
Multiplier2× all purchases
Budget₹8.2L
📝 Create Custom Campaign
Design a tailored retention strategy for a specific customer segment
Churn Rate by Region
East region highest at 57.8%
Satisfaction vs Churn
Churned vs Retained by score
Churn by Tenure
Long-tenure customers churning most
Age Group Analysis
Churn % across age brackets
📄 Export Reports
🤖 Churn Prediction Engine

Enter customer features to calculate churn probability.
Runs the real trained Random Forest in your browser — no server required.

🔌 Flask API — Already Integrated!
The prediction engine calls your local model automatically. Start the server with: 
python app.py
# → Running on http://localhost:5001
# Falls back to rule-based model if Flask is offline
⚙️ Dashboard Configuration
Model Integration
POST endpoint that accepts customer features and returns {"probability": 0.87}
Risk Thresholds
Display