Consumer Subscription Behavior Prediction using Machine Learning

Abstract

Nowadays the e-commerce market is becoming increasingly competitive in the era of industry 4.0, which highlights the importance of customer retention strategies, especially through consumer subscription rates. In order to survive in the competitive environment, companies should understand the purchasing behavior of users, especially in the subscription model business as it is becoming one of the most important elements of revenue generation in business. This study utilized six supervised machine learning algorithms namely K-Nearest Neighbor, Support Vector Machine, Logistic Regression, Decision Tree, Random Forest and Extreme Gradient Boosting in predicting whether consumers will choose a subscription service. Among the six supervised machine learning, the champion model found which shows that the Random Forest model has the highest accuracy. It is also found that promo codes used, discount applied, and previous purchases are the most influential features for customers to choose to subscribe. This will enable subscription service providers to revise their subscription plan in order to attract more customers.