Research on the application of cluster analysis in cross-border e-commerce customer segmentation and market strategy

Abstract

E-commerce platforms are becoming increasingly competitive, and how to attract and retain users has become a problem that operators need to solve. The study improves the K-means clustering algorithm by introducing the concepts of mean sample distance and SSE and using the median to construct profile coefficients. Based on this, the user labels of cross-border e-commerce companies are selected, and clustering segmentation is performed after collecting the user data of Company T on cross-border e-commerce platforms to reveal the basic characteristics, purchasing preferences, and behavioral habits of different e-commerce users. A cross-border e-commerce marketing strategy based on customer segmentation has been formulated, and its effectiveness has been verified. Through clustering analysis, users are successfully classified into three types: high-value (4.01%), dynamic (7.35%), and growth (88.64%). After implementing the marketplace strategy, the sample companies gained significant growth in user growth, referral effect, and merchandise sales, as reflected in the number of first-buy users (63.46%), effective users (21.45%), ad click rate (18%~22%) and sales volume (166.43%). This study can help cross-border e-commerce enterprises better understand user needs and behavioral patterns and formulate more accurate marketing strategies to enhance their competitiveness and market share.

Keywords

• k-means
• profile coefficient
• cluster analysis
• customer segmentation
• marketing strategy
• 22a10