Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/318810 
Year of Publication: 
2023
Citation: 
[Journal:] Business Systems Research (BSR) [ISSN:] 1847-9375 [Volume:] 14 [Issue:] 1 [Year:] 2023 [Pages:] 26-53
Publisher: 
Sciendo, Warsaw
Abstract: 
Background: Customer segmentation has become one of the most innovative ways which help businesses adopt appropriate marketing campaigns and reach targeted customers. The RFM model and machine learning combination have been widely applied in various areas. Motivations: With the rapid increase of transactional data, the RFM model can accurately segment customers and provide deeper insights into customers' purchasing behaviour. However, the traditional RFM model is limited to 3 variables, Recency, Frequency and Monetary, without revealing segments based on demographic features. Meanwhile, the contribution of demographic characteristics to marketing strategies is extremely important. Methods/Approach: The article proposed an extended RFMD model (D-Demographic) with a combination of behavioural and demographic variables. Customer segmentation can be performed effectively using the RFMD model, K-Means, and K-Prototype algorithms. Results: The extended model is applied to the retail dataset, and the experimental result shows 5 clusters with different features. The effectiveness of the new model is measured by the Adjusted Rand Index and Adjusted Mutual Information. Furthermore, we use Cohort analysis to analyse customer retention rates and recommend marketing strategies for each segment. Conclusions: According to the evaluation, the proposed RMFD model was deployed with stable results created by two clustering algorithms. Businesses can apply this model to deeply understand customer behaviour with their demographics and launch efficient campaigns.
Subjects: 
Customer segmentation
RFMD model
K-Means
One hot encoding
K-Prototypes
Cohort analysis
machine learning
JEL: 
C61
C63
C67
Persistent Identifier of the first edition: 
Creative Commons License: 
cc-by Logo
Document Type: 
Article

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.