FICO credit scores have a pervasive influence on the financial well-being of U.S. consumers. However, consumers have limited knowledge of what economic activities are considered in the scoring algorithm and how these activities affect their credit scores. A lack of understanding about the algorithm is detrimental to consumer welfare since consumers will be unable to improve their credit scores even if they have the means and resources to do so. To address this, we used interpretable machine learning to analyze a large volume of credit report data to identify the relationships between financial behaviors and credit scores. Our approach involved calibrating the best-performing machine learning model and providing interpretability through the Shapley Additive exPlanations (SHAP) procedure from cooperative game theory. This combined approach provides actionable insights with useful implications to help consumers manage their credit scores more effectively. Our study contributes to transformative consumer research literature by improving the quality of consumers' decisions and enhancing consumer welfare.