Predicting the Success of Marketing Campaigns in the Banking Sector

Abstract

In today’s competitive banking environment, identifying potential customers who are most likely to respond positively to marketing campaigns is essential for optimizing resource allocation and improving conversion rates. This study aims to predict the success of marketing campaigns in the banking sector by analysing the UCI Bank Marketing Dataset, which contains demographic and behavioural attributes of customers. The dependent variable is whether a client subscribes to a term deposit, while independent variables include age, job type, marital status, education, contact frequency, and previous campaign outcomes. Using Logistic Regression, Decision Tree, and Random Forest models, the research evaluates predictive accuracy and identifies the most influential factors driving campaign success. The data were preprocesses through encoding and standardization, followed by exploratory data analysis (EDA) and model evaluation using metrics such as Accuracy, ROC-AUC, and Classification Reports. Results reveal that the Random Forest classifier achieved the highest accuracy (90.6%) and identified duration of contact and previous campaign outcomes as the strongest predictors. The study contributes to better targeting strategies and customer segmentation for financial institutions, enabling data-driven decision-making in marketing.