This study explored whether combining promotional strategies with specific shipping types influenced consumer spending. The assumption was that a discount paired with fast or free shipping might have increased purchase amounts. Using a real-world dataset that included demographic variables, purchase history, promo code use, and shipping preferences, the outcome variable, purchase amount, was treated as continuous.Linear regression and XGBoost models were used to test both main and interaction effects. SHAP values and supporting visuals helped interpret model behavior.Across all models, the findings were consistent: there was no statistically significant relationship between promo code use, shipping type, or their interaction and how much people spent. SHAP confirmed that these features had little predictive weight.Although the hypothesis was not supported, the results raised broader questions. Some of the most widely used marketing tactics may not have always moved the needle on spending. More detailed, context-rich data might have been needed to uncover what actually did.
Bente Beelen (Tue,) studied this question.