Network-Based Analysis of Heterogeneous Consideration-Then-Choice Customer Preferences with Market Segmentations

Network-based analyses have been shown to be effective in understanding customer preferences by modeling the interactions and relationships between customers and products as a complex network system. Certain network representations, such as bipartite networks, can capture customers’ two-stage (consideration-then-choice) decision-making processes by constructing either the “consideration” or “choice” links between the customer nodes and the product nodes. However, there is a dearth of research that examines network-based approaches to understanding complex heterogeneous customer preferences and the influence of product features in different market segments. This paper presents a market-segmented network modeling approach for understanding heterogeneous customer preferences in two-stage (consideration-then-choice) decision-making. Joint correspondence analysis is utilized to identify the correlations between product association networks and customer attributes, and then market segments are characterized by clustering customer attributes. We then construct bipartite customer–product networks and use the Exponential Random Graph Model to investigate factors that influence customer decision-making processes and how they vary among customer groups. A case study using real customer survey data for vacuum cleaners, a common household appliance with various product categories and a sizable market, serves to demonstrate the approach. The survey has been systematically designed and conducted on Cint platform to collect customer considerations and choices, product features, and customer attributes. Our findings reveal that customers are heterogeneous across different market segments, which can be clustered based on their demographic attributes, usage contexts, and personal viewpoints. Within the identified market segments based on these aforementioned customer attributes, customer preferences toward product attributes show heterogeneity in different stages of choice-making. Particularly, it is observed that the majority of the product design attributes receive more attention in the consideration stage than in the choice stage. Our study advances the use of network-based models for analyzing customer preference heterogeneity across different market segments and in different stages of customers’ decision-making.