Customer segmentation is one of the most salient factors to be considered in the making of data-driven decisions about the consumers for facilitating their engagement and retention. This study offers general insights regarding promising clustering of customers with K-Means- and GMM-based methods using realistic transaction data from an e-commerce platform. Various segments of customers were created depending on a few vital behavioral characteristics such as frequency of purchase, average order value, probability of churn, and most preferred times for shopping. The segmentation is, then, classified by a Logistic Regression, Decision Tree, Random Forest, SVM, XGBoost, Gradient Boost, and LightGBM to predict customer segments. The models’ effectiveness in classifying segments is judged by accuracy, precision, recall, and F1 score. This, then, builds up into giving insights into customer behavior in an attempt to help the companies design data-driven marketing efforts to maximize retention and profits in the highly competitive e-commerce market.