# ML & Propensity Models

Canonical URL: https://babadook.tech/modules/ml-propensity-models
Entity: modules
Content hash: 0b84e2e9962feae657ce39a857f45f2aecbc620ef1b9948c3d2f31bb3187dd73

Pre-built machine learning models for propensity-to-convert, churn prediction, lifetime value estimation, and next-best-action recommendation.

Pre-built machine learning models for propensity-to-convert, churn prediction, lifetime value estimation, and next-best-action recommendation.

## predict,-don't-react

Predict, Don't React ML & Propensity Models continuously score every user on their likelihood of key outcomes — converting, churning, expanding, or engaging with specific content. These scores update in real-time as new behavioral data streams in. Propensity scores feed directly into DXP CDP for segmentation and DXP Marketing Engine for targeting. High churn-risk users automatically receive retention campaigns. High-propensity converters see accelerated conversion flows.

### data Viz

## ML Model Specifications

- **Models:** Propensity-to-convert, churn risk, LTV, next-best-action
- **Training:** Continuous retraining with new behavioral data
- **Scoring:** Real-time scoring on event stream
- **Features:** Auto-generated from behavioral events + custom features
- **Explainability:** SHAP values for model interpretability

## features

ltv: true

churn prediction: true

next best action: true

propensity convert: true

## Relationships

### products

- DXP — https://babadook.tech/products/dxp

### capabilities

- Advanced Analytics — https://babadook.tech/capabilities/advanced-analytics
- Customer Intelligence — https://babadook.tech/capabilities/customer-intelligence

### sub modules

- Analytics & Intelligence — https://babadook.tech/sub-modules/analytics-intelligence

### use cases

- Behavioral Analytics
- Customer Data Unification