Data-Driven Prediction-Making on customer churn in a circular economy through RFM and clustering algorithms

Enterprises are facing the two-sided challenge of adapting from one aspect to the needs of the new emerging circular economy requirements and the tough competition deriving from new emerging digital business models, hence they must develop innovative strategies to identify customers and serve them based on their behavior and characteristics. Identification of customer clusters serves as a basis for an appropriate strategy to predict customer churn. There are many different models and algorithms to use for customer segmentation, each with specific characteristics. Some of these models and two of many frameworks that use a combination of the algorithms are explained in the second section of this paper. The methodology is explained in section three, represented by a survey in the coffee industry. The purpose of the survey is to collect customer data based on predefined variables, which will serve as additional information on the RFM analysis. On the last section we have concluded with a proposition of a new integrated method for customer clustering, which will be part of our future work.