Title
Segmenting market structure from multi-channel clickstream data: a novel generative model
Abstract
Competitive analysis has long been recognized as the cornerstones of firm’s strategic management and business activities. With the advent of the multi-channel clickstream, this paper studies the competitive market structure by developing a novel generative model. We first aggregate the multi-channel clickstream data to construct a consideration set for each user. Then, a novel sparse influence topic model (SITM) is proposed to segment an overall market into submarkets by leveraging the consideration sets at the individual level. Compared with the current generative models, the proposed SITM model considers the limited interest and the influence of products to generate users’ choice behaviors. Based on the multi-channel clickstream data from 109,081 users on 3779 cars, we empirically analyze the competition structure in China’s automotive market. Experimental results show that the proposed model can obtain deep insights of the competitive market structure and the competition power of each car in the market. It can also help managers understand user’s personalized interesting in the competitive market.
Year
DOI
Venue
2020
10.1007/s10660-019-09393-0
Electronic Commerce Research
Keywords
DocType
Volume
Clickstream data, Market structure, Generative model, Personal interest, Product influence
Journal
20
Issue
ISSN
Citations 
3
1389-5753
0
PageRank 
References 
Authors
0.34
1
5
Name
Order
Citations
PageRank
Yang Qian100.68
Yuanchun Jiang218421.24
Yanan Du300.34
Jianshan Sun419217.65
Yezheng Liu514524.69