Skip to main content

Reactivation

Blair Matheson

← Back to all Predictive Models

PREDICTIVE
🎯 Purpose & Use-Cases

This model is designed to predict the probability that a churned customer will reactivate and return to active status within the next 2 iterations.

Win-Back Campaigns
Target churned customers with the highest likelihood to return

Strategic Resource Allocation
Enable cost-effective reactivation to:

  • Optimize promotional budgets by targeting receptive churned customers
  • Customize win-back offers based on reactivation probability
  • Focus retention efforts on customers most likely to respond
🔍 Model Scope
  • Population Lifecycle Stages: All customers in "Churn" lifecycle stage
🔄 Update Frequency
  • Training: Runs every iteration (unless data drift/changes require retraining)
  • Inference: Runs daily in the daily process
📊 Output

The following outputs are stored in the customer profiles (as internal fields):

  • Reactivation Probability Score: Probability of this customer to be reactivated within the next 2 iterations
  • Percentile rank in Reactivation Probability: Percentile rank among all churned customers by Reactivation Probability Score.
    On a scale of 1 (lowest) to 100 (highest)
  • Permille rank in Reactivation Probability: Permille rank among all churned customers by Reactivation Probability Score.
    On a scale of 1 (lowest) to 1000 (highest), providing higher-resolution segmentation than the percentile ranking
⚙️ Model Specifications
  • Training Window: default = 20 iterations
  • Prediction Horizon: default = 2 iterations
🧠 How It Works
  • Catboost Classifier, which is the state-of-the-art model for tabular data
  • Individual models per lifecycle stage for optimal accuracy
  • Hyperparameters tuning in the first model training for optimal model parameters
✨ Data Quality & Automation

Data Cleaning & Imputation: Handles missing values, removes constant features, and groups rare categorical values

Feature Engineering & Encoding: Converts categorical variables to numerical format using target-based encoding and applies scaling transformations

Feature Selection & Optimization: Removes low-variance and highly correlated features, selecting the most predictive variables

📁 Data Requirements
  • Historical customer profiles (customer activity data, demographic data, campaign history, etc.)
  • Minimum 3 iterations of historical data required for training (1 iteration for feature generation and 2 iterations for target definition)
  • Minimum number of 500 customers per lifecycle stage required for training
  • Minimum number of 50 customers per lifecycle stage that reactivated in the past required for training

Related articles