چکیده:
Conventional segmentation methods are designed based solely on the three components of recency (R), frequency (F), and monetary value (M) and do not consider changes in customer behavior over time. Therefore, to achieve optimal segmentation, the purpose of this research is to apply a set of statistical calculations such as slope and derivative with respect to time, and data mining methods such as K-means and self-organizing maps (SOM) to define new variables in order to investigate the trend of changes in customer purchasing behavior. The results of the research show that considering the slope of changes in customer behavior (R, F, and M) and giving more value to recent behaviors compared to past behaviors in customer segmentation leads to increased accuracy in predicting future behavior and greater homogeneity of customers in each segment. Based on the proposed method, customers were categorized into four segments: best, spender, repeater, and lost, and each segment was further classified into two sub-segments: ascending and descending, in order to better and more accurately understand customers based on how their behavior changes. Finally, while describing the characteristics of each segment and sub-segment, appropriate strategies for managing their customers are presented.
خلاصه ماشینی:
Therefore, to achieve optimal segmentation, the purpose of this research is to employ a set of statistical calculations such as slope and derivative with respect to time, and data mining methods such as K-means and Self-Organizing Maps (SOM) to define new variables in order to investigate the trend of changes in customer purchasing behavior.
Customer segmentation is one of the data mining methods that aims to group customers with similar needs and purchasing behavior to determine different strategies in order to maximize response to marketing programs and reduce organizational costs (Smith, 1956; Azizi, Hosseinabadi & Belaghi Inanlu, 1393).
Recency, Frequency, and Monetary 1 Therefore, the aim of the present research is to employ a set of data mining methods and define some new variables in the RFM method, in order to provide innovative methods for appropriate and effective customer segmentation; so that businesses can, with the help of these methods, differentiate between customers based on their purchasing behavioral changes and also identify customers who are at risk of churning or becoming unprofitable customers, and consequently, deal with customers with different behaviors in different ways.
The characteristics of the RFM model (recency, frequency, and monetary value of purchase) are a suitable method for segmenting customers based on analyzing their purchasing behavior and have been successfully used in various studies (Sohrabi & Khanlari, 2007; Razmi & Ghanbari, 1388; Kosment & Debak, 2013; Akhondzadeh Noghabi et al.
(
, 
, 
)
