Gaining insight to B2B relationships through new segmentation approaches: Not all relationships are equal

Matthew O'Brien,Ying Liu,Hongyu Chen,Robert Lusch

DOI: https://doi.org/10.1016/j.eswa.2020.113767

IF: 8.5

2020-12-01

Expert Systems with Applications

Abstract:<p>B2B market segmentation has both structure complexity and computation complexity. The existing market segmentation methods can not directly address these two challenges simultaneously and provide a comprehensive view of the whole problem. Nor are they able to provide guidance on the selection of the most suitable solution among candidates. This study formulates the B2B segmentation as a multi-dimensional optimization problem that integrates both customer behavior and marketing effectiveness. It applies an integrated segmentation method that unifies two market segmentation approaches: the Embedded Exchange Approach (a descriptive model) and the Predictive Satisfaction Approach (a predictive model). It proposes the use of an evolutionary based, multi-objective segmentation method to solve the structural and computational challenges. The method generates a set of Pareto optimal solutions which not only gives a holistic view of possible solutions in the Pareto optimal space but also allows marketers to use solution selection algorithm based on the properties of Pareto optimal sets. The study develops a solution selection algorithm that represents a good tradeoff of two objectives based on the geometric shape of the Pareto optimal solution front.</p>

computer science, artificial intelligence,engineering, electrical & electronic,operations research & management science

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the structural complexity and computational complexity existing in B2B market segmentation. The existing market segmentation methods cannot directly address these two challenges simultaneously and provide a comprehensive view of the entire problem. Moreover, these methods are also unable to provide guidance for choosing the most appropriate solution among the candidate solutions. Therefore, this paper proposes a new multi - dimensional optimization problem framework, aiming to integrate customer behavior and marketing effectiveness. By combining the descriptive model (embedded exchange method) and the predictive model (predictive satisfaction method), and using the evolution - based multi - objective segmentation method to solve these problems. This method can not only generate a set of Pareto - optimal solutions, thus providing a comprehensive view of possible solutions, but also allows marketers to choose solutions based on the characteristics of the Pareto - optimal set. In addition, the research has also developed a solution selection algorithm, which finds a good compromise between the two objectives according to the geometry of the Pareto - optimal solution frontier.