Deposits are an important source of funding that enables commercial banks to maintain liquidity and operate prudently, while telemarketing is a major channel through which banks promote term deposit products. A key challenge in precision marketing is to identify prospective subscribers while reducing ineffective contacts and to translate predictions into customer relationship management actions. Using the UCI Bank Marketing dataset, this study develops a five-stage framework that integrates model comparison, class imbalance treatment, feature selection, SHAP interpretation, and customer relationship management. By comparing seven types of machine learning models and multiple sampling methods and ratios, the study identifies the optimal Optuna-tuned ROS-LightGBM model, which achieves an 88.78% recall and a 93.17% ROC-AUC on a test set that retains the original class distribution. SHAP analysis shows that marketing interaction features account for 74.31% of the total contribution, with call duration, contact channel, previous campaign outcome, and month exerting the strongest effects. This study further combines predicted probabilities with customer balances to classify customers into four groups: priority conversion, potential cultivation, relationship maintenance, and low-cost maintenance. It then proposes differentiated contact and maintenance strategies, establishing a practical pathway from prospective customer identification and prediction interpretation to customer relationship management, and providing a reference for the refined management of bank term deposit telemarketing.

Yuxuan Chen, Exiang Luo. ROS-LightGBM-Based Prediction of Term Deposit Subscriptions in Bank Telemarketing and CRM Strategy Analysis. Academic Journal of Business & Management (2026), Vol. 8, Issue 7: 99-117. https://doi.org/10.25236/AJBM.2026.080714.

[1] Gatev E, Schuermann T, Strahan P E. Managing bank liquidity risk: How deposit-loan synergies vary with market conditions[J]. The Review of Financial Studies, 2009, 22(3): 995-1020.

[2] Cornett M M, McNutt J J, Strahan P E, et al. Liquidity risk management and credit supply in the financial crisis[J]. Journal of Financial Economics, 2011, 101(2): 297-312.

[3] Ivashina V, Scharfstein D. Bank Lending During the Financial Crisis of 2008[J]. Journal of Financial Economics, 2010, 97(3): 319-338.

[4] Egan M, Hortaçsu A, Matvos G. Deposit competition and financial fragility: Evidence from the US banking sector[J]. American Economic Review, 2017, 107(1): 169-216.

[5] Clemes M D, Gan C, Zhang D. Customer switching behaviour in the Chinese retail banking industry[J]. International Journal of Bank Marketing, 2010, 28(7): 519-546.

[6] Liao S H, Chen Y J, Hsieh H H. Mining Customer Knowledge for Direct Selling and Marketing[J]. Expert Systems with Applications, 2011, 38(5): 6059-6069.

[7] Moro S, Cortez P, Rita P. A Data-Driven Approach to Predict the Success of Bank Telemarketing[J]. Decision Support Systems, 2014, 62: 22-31.

[8] Feng Y, Yin Y, Wang D, et al. A Dynamic Ensemble Selection Method for Bank Telemarketing Sales Prediction[J]. Journal of Business Research, 2022, 139: 368-382.

[9] Xie C, Zhang J L, Zhu Y, et al. How to Improve the Success of Bank Telemarketing? Prediction and Interpretability Analysis Based on Machine Learning[J]. Computers & Industrial Engineering, 2023, 175: 108874.

[10] Moro S, Cortez P, Rita P. Using customer lifetime value and neural networks to improve the prediction of bank deposit subscription in telemarketing campaigns[J]. Neural Computing and Applications, 2015, 26(1): 131-139.

[11] Li S, Sun B, Montgomery A L. Cross-selling the right product to the right customer at the right time[J]. Journal of Marketing Research, 2011, 48(4): 683-700.

[12] Godfrey A, Seiders K, Voss G B. Enough is enough! The fine line in executing multichannel relational communication[J]. Journal of Marketing, 2011, 75(4): 94-109.

[13] Ngai E W T, Xiu L, Chau D C K. Application of Data Mining Techniques in Customer Relationship Management: A Literature Review and Classification[J]. Expert Systems with Applications, 2009, 36(2): 2592-2602.

[14] Payne A, Frow P. A Strategic Framework for Customer Relationship Management[J]. Journal of Marketing, 2005, 69(4): 167-176.

[15] Koumétio Tékouabou S C, Gherghina S C, Toulni H, et al. A Machine Learning Framework towards Bank Telemarketing Prediction[J]. Journal of Risk and Financial Management, 2022, 15(6): 269.

[16] Guo W, Yao Y, Liu L, et al. A Novel Ensemble Approach for Estimating the Competency of Bank Telemarketing[J]. Scientific Reports, 2023, 13: 20819.

[17] Dong B, Zhang X, Liu Y, et al. Risk-Sensitive Machine Learning for Financial Decision Modeling Under Imbalanced Data: Evidence from Bank Telemarketing[J]. Entropy, 2026, 28(3): 354.

[18] Nasir F, Ahmed A A, Yevseyeva I, et al. Marketing Analytics in Banking 4.0: A Two-Stage Explainable AI Framework for High-Accuracy and Well-Calibrated Predictions[J]. PLOS One, 2026, 21(5): e0348767.

[19] Chawla N V, Bowyer K W, Hall L O, et al. SMOTE: Synthetic Minority Over-Sampling Technique[J]. Journal of Artificial Intelligence Research, 2002, 16: 321-357.

[20] Ke G, Meng Q, Finley T, et al. LightGBM: A Highly Efficient Gradient Boosting Decision Tree[C]. Advances in Neural Information Processing Systems, 2017, 30: 3146-3154.

[21] Lundberg S M, Erion G, Chen H, et al. From Local Explanations to Global Understanding with Explainable AI for Trees[J]. Nature Machine Intelligence, 2020, 2: 56-67.

[22] Peppers D, Rogers M. Managing Customer Relationships: A Strategic Framework[M]. 3rd ed. Hoboken, NJ: John Wiley & Sons, 2016.

[23] Kumar S, Kumar A B, Ravi V. Explainable Artificial Intelligence for Analytical Customer Relationship Management in Banking and Finance[C]. In: Intelligent Systems Design and Applications. Springer, 2024: 101-112.

[24] Apicella A, Isgrò F, Prevete R. Don't Push the Button! Exploring Data Leakage Risks in Machine Learning and Transfer Learning[J]. Artificial Intelligence Review, 2025, 58: 339.

[25] Sasse L, Nicolaisen-Sobesky E, Dukart J, et al. Overview of Leakage Scenarios in Supervised Machine Learning[J]. Journal of Big Data, 2025, 12: 135.