Skip to main content

Future Value

Blair Matheson
  • Updated

← Back to all Predictive Models

PREDICTIVE
🎯 Purpose & Use-Cases

This model is designed to predict the future value of a customer according to the site's main KPI (Total Revenue, Total Order, etc.).

Revenue Optimization
Allocate marketing budgets based on predicted customer value

Target Group Marketing
Enable strategic customer targeting to:

  • Prioritize high-value customers for premium campaigns
  • Customize offer levels based on predicted spending capacity
  • Focus retention efforts on highest value customers
🔍 Model Scope
  • Population Lifecycle Stages: Customers in all lifecycle stages besides "Dormant"
  • Target: Next year aggregated KPI for each customer in the population lifecycle stages
🔄 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):

  • Future Value Score: Future value prediction of customer over the next year
  • Percentile rank in Future Value: Percentile rank among all customers by Future Value.
    On a scale of 1 (lowest) to 100 (highest)
  • Permille rank in Future Value: Permille rank among all customers by Future Value.
    On a scale of 1 (lowest) to 1000 (highest), providing higher-resolution segmentation than the percentile ranking
  • Percentile rank in LCS Future Value: Percentile rank among same lifecycle stage customers by Future Value.
    On a scale of 1 (lowest) to 100 (highest)
  • Permille rank in LCS Future Value: Permille rank among same lifecycle stage customers by Future Value.
    On a scale of 1 (lowest) to 1000 (highest), providing higher-resolution segmentation than the percentile ranking
⚙️ Model Specifications
  • Training Window: default = One Year
  • Prediction Horizon: default = One Year
🧠 How It Works
  • Catboost Regressor, 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 1 year of historical data required for training
  • Minimum number of 1000 customers per lifecycle stage required for training

Related articles