Designing customer segmentation model for analysing consumer data: Case: Consumer segmentation model for retail sales

Abstract

Customer segmentation has a crucial impact on today’s highly competitive business environment, and it is especially affecting to organizations marketing processes. The main idea behind customer segmentation is to divide customers into homogeneous groups based on their various characteristics. This enables organizations to tailor their marketing strategies to the various needs of different customers, without creating distinct plan for each individual customer. Customer segmentation can also help organizations to understand their customers better and to find latent business opportunities. There are many different approaches to customer segmentation in the consumer markets, but the four major approaches are behavioural, demographic, geographic, and psychographic segmentation. To utilize these approaches, it is recommended to use machine calculation and various data scientific methods to be able to process bigger amounts of data.

The objective of this research was to design customer segmentation model, that is appropriate for analysing quantitative sales data in retailer consumer market. The purpose of the model is to show how behavioural based consumer segmentation could be implemented and utilized in the client organization. Research includes the theory and the empirical case study section, which follows the design science research as a strategic framework. The study explores various consumer segmentation approaches and data scientific methods that can be utilized in the segmentation model. The model is developed in case study, in which the identified methods are practically applied by using the client organization’s consumer sales data. The study is also investigating data scientific frameworks that can be utilized in the iterative development process of segmentation model along with the design science.

The main result of this research is the consumer behaviour-based segmentation model, that utilizes customer’s recency, frequency, and monetary based modelling as a base segmentation method, which is widely used in the behavioural segmentation. The segmentation model divides consumers into seven homogenous segments based on their buying behaviour during the last five years. The used method is easy to understand, and it enables arbitrary tailoring of the limit values and the labels of segments. Some additional geographic and product related attributes were also added to model as an explanatory features. Another segmentation method considered in this research is K-Means clustering. The study found that this unsupervised method would be a proper solution if more than three features were used as a dividing criterion in the segmentation. However, clustering is not completely excluded from the model, as it offers a good comparison for manually created segments and enables several further development opportunities.