A Machine Learning Approach to Online Market Segmentation

Abstract

This paper proposes a new approach for online market segmentation using clickstream data. The concept of the customer journey is used to frame the analysis of the online search and buying behaviour. A novel methodology that employs Machine Learning (ML) is proposed and a detailed framework including data pre-processing and model training shows it can be applied to online market segmentation based on clickstream data, which is a direct measure of user behaviour. This inductive approach can help us generates distinctive online market segments that have face validity to practising managers, and can be applied at scale to travel, e-commerce and more generally B2C websites. The distinct market segments that are identified can be grouped according to distinctive search and buying behaviours, which are statistically and behaviourally homogenous within market segments and heterogeneous between market segments. The theory contribution to marketing is described and the managerial implications of the results are outlined.