Skip to main content

Propensity Models

Once the data are well understood, modeling will begin. If there is a stable and robust response variable present, then a Logistic Regression or other kind of model is a logical next step, assigning a loyalty or lifetime value score to each record.Depending on the modeling objective and the available response variable(s), modeling efforts may result in:

  • Response Models
  • Revenue/Profit Models
  • Retention/Loyalty/Persistency Models
  • Lifetime Value Models
  • In the event where a Marketing history is not available for a given Product or Product Class, latent variable analysis (an extension of exploratory analysis) can also be used to generate a propensity model