Which statistical models do you prefer when predicting user churn or conversion probability for Credible?
