Internet Customer Segmentation Using Web Log Data

Abstract

The objective of this paper is to analyze web transaction log data that reveal customer behavior in the Internet channel, and to provide a useful online customer segmentation scheme. To achieve this, we analyze the relationship between the behavior of customers for online pet shops and revenue. We use the decision-tree method as a data-mining technique, and clustering analysis to segment customers. We perform the study in two stages. First, we investigate the web transaction data of both the member customers and nonmember customers of a Korean online pet shop. Second, we narrow down the study focus and analyze only the member customers’ demographic data and their web transaction data. As a result, we obtain several meaningful segments based on customers’ transaction behavior and demographic characteristics. We use web log data to analyze customer transaction behavior and log-in information to analyze customer demographic characteristics. We discuss some strategic implications, for online shopping mall marketing, suggested by the acquired market segments.