Analysis on E-commerce Order Cancellations Using Market Segmentation Approach

Abstract

This study investigates the application of market segmentation on E-commerce canceled orders. It uses a transnational dataset that contains transactions of an online retail store during a year. The analysis process includes 1) an exploratory data analysis on the canceled orders which makes up a considerably amount of the dataset to show their characteristics. 2) a production segmentation that utilize the k-means clustering to create 5 product clusters. 3) a customer segmentation with k-means clustering using the production segments and customer features which results in 7 segments. In the process, the study compares silhouette scores and applies principal component analysis to optimize the number of clusters. The conclusion shows that market segmentation serves as an effective tool to distinguish products and consumers with different characteristics and help make suggestions to businesses. Also, including attitudinal features into the analysis process will result in improved customer profiles.