A dynamic customer segmentation approach by combining LRFMS and multivariate time series clustering.

Area of Science:

• Business Analytics
• Data Mining
• Marketing Science

Background:

• Effective customer analysis and management are crucial for automotive parts agents in the Industrial Internet era.
• Dynamic customer segmentation aids in identifying distinct customer groups.
• Existing methods like RFM and univariate time series clustering have limitations.

Purpose of the Study:

• To propose an improved dynamic customer segmentation approach.
• To address limitations of the traditional RFM model and univariate clustering.
• To enhance customer analysis for automotive parts marketing.

Main Methods:

• Combining Length, Recency, Frequency, Monetary, and Satisfaction (LRFMS) variables.
• Utilizing multivariate time series clustering algorithms.
• Applying distance measurement methods: DTW-D, SBD, and CID.

Main Results:

• The proposed LRFMS and multivariate time series clustering approach effectively segments automotive parts customers.
• Empirical analysis validates the approach's effectiveness compared to existing methods.
• Identified distinct customer segments with actionable marketing insights.

Conclusions:

• The LRFMS and multivariate time series clustering method offers a robust solution for dynamic customer segmentation.
• This approach provides valuable insights for targeted marketing strategies in the automotive parts industry.
• Enhanced customer understanding leads to improved marketing performance.