How Lens360 Works
A transparent look at our ML pipeline — from raw telecom data to actionable business recommendations with measurable ROI.
The Data Journey
How customer data flows through our system to produce actionable intelligence
Raw Data
CDRs, billing, CRM, usage logs
Feature Engineering
50+ behavioral signals extracted
Model Pipeline
6 specialized ML models
Pattern Detection
7 behavioral segments
Business Output
Prioritized actions + ROI
Data Streams
Usage Data
- Traffic volumes
- VAS consumption
- Data usage
- Voice minutes
Revenue Data
- Monthly ARPU
- PAYG spend
- Bolton purchases
- Payment history
Product Data
- Current plans
- Add-ons active
- Upgrade history
- Eligibility
Engagement Data
- Tenure
- Support tickets
- App usage
- Channel preference
Model Architecture
Our hybrid system uses multiple specialized models, each optimized for specific tasks
Model Deep Dives
Click each model to explore its features, importance weights, and technical implementation
Churn Prediction
Predicts the probability that a customer will churn within the next 90 days. Uses ensemble learning to identify complex patterns in customer behavior that precede churn events.
Key Features & Importance
Used For
Collaborative Filtering
Discovers latent patterns in customer-product interactions. Finds customers with similar behaviors and recommends products that similar customers have adopted.
Key Features & Importance
Used For
Content-Based Filtering
Recommends products based on customer attributes and product characteristics. Matches customer profiles to ideal product features using cosine similarity.
Key Features & Importance
Used For
Pattern Detection
Classifies customers into 7 actionable behavioral patterns using domain-expert rules. Each pattern maps directly to a business action and data product.
Key Features & Importance
Used For
Hybrid Engine
Combines content-based (60%) and collaborative filtering (40%) scores with business rules overlay. Ensures recommendations are both personalized and commercially viable.
Key Features & Importance
Used For
Acquisition Scoring
Scores prospects for acquisition safety — predicting which new customers are likely to stay vs. churn quickly. Optimizes CAC by targeting high-retention prospects.
Key Features & Importance
Used For
Platform Integration
How each model connects to the business outcomes you see in the platform
Why This Architecture?
Explainable
Every recommendation comes with clear reasoning — no black box decisions.
Measurable
ROI projections for every action. Know the expected impact before you act.
Actionable
Patterns map directly to business actions. No interpretation needed.