Endy Wulan Agustin, Kurnia Uthami, Arvan Izzatul Ulfa, Lusiana Efrizoni, Rahmaddeni Rahmaddeni

Optimization of Customer Segmentation in the Retail Industry Using the K-Medoid Algorithm

Introduction

Optimization of customer segmentation in the retail industry using the k-medoid algorithm. Optimize retail customer segmentation using the K-Medoid algorithm. Enhance marketing strategies and loyalty programs by achieving more accurate, homogeneous customer clusters from big data.

Abstract

The retail industry faces significant challenges in understanding increasingly complex customer behavior due to massive data growth. One major obstacle is suboptimal customer segmentation, leading to ineffective marketing strategies. This study aims to optimize customer segmentation by implementing the K-Medoid algorithm, which excels in handling outliers and producing more stable clusters compared to K-Means. The dataset consists of over 10,000 customer transactions from a major retail company in Indonesia. The research process includes data collection and preprocessing, K-Medoid algorithm implementation, and performance evaluation using the silhouette score. The results indicate that the K-Medoid algorithm achieves more accurate customer segmentation, with a silhouette score of 0.39. The generated clusters exhibit greater homogeneity, enabling companies to design more targeted marketing strategies, such as specific discount offers and tailored loyalty programs. Based on these findings, the K-Medoid algorithm is recommended to enhance customer management effectiveness in the retail industry. This study contributes to selecting a more suitable algorithm for customer segmentation in the era of big data and opens opportunities for further exploration of hybrid algorithms and additional evaluation metrics.

Review

This study, titled 'Optimization of Customer Segmentation in the Retail Industry Using the K-Medoid Algorithm,' addresses a highly pertinent challenge within the modern retail landscape: effectively understanding complex customer behavior amidst massive data growth. The paper proposes the implementation of the K-Medoid algorithm as a solution to suboptimal customer segmentation, aiming to enable more effective marketing strategies. Utilizing a substantial dataset of over 10,000 customer transactions from a major Indonesian retailer, the research outlines a clear methodological approach encompassing data preprocessing, algorithm application, and performance evaluation via the silhouette score. The premise of using K-Medoid for its robustness against outliers and stability makes a strong case for its application in real-world retail data. A key strength of this research lies in its direct relevance and practical implications for the retail industry. The selection of the K-Medoid algorithm is well-justified by its inherent advantages over K-Means, particularly in handling noisy transactional data, which is common in retail environments. The reported outcome of a silhouette score of 0.39, while moderate, is stated to lead to "more accurate" segmentation and "greater homogeneity" within clusters, providing a tangible basis for developing targeted marketing efforts such as specific discount offers and tailored loyalty programs. The use of real-world data from a significant regional player adds credibility and immediately translates the findings into actionable business intelligence, contributing meaningfully to the discourse on selecting appropriate algorithms for customer management in the big data era. While the study presents a valuable contribution, there are a few areas that could benefit from further elaboration or critical discussion in the full paper. Firstly, while a silhouette score of 0.39 is achieved, the interpretation of its "accuracy" and "homogeneity" could be contextualized more thoroughly against industry benchmarks or comparative studies, especially given that such a score is often considered fair rather than exceptionally strong. It would also strengthen the claims if the paper explicitly detailed a comparative analysis with the K-Means algorithm, beyond the theoretical advantages, to empirically demonstrate the K-Medoid's superior performance on this specific dataset. Furthermore, clarifying the methodology for determining the optimal number of clusters (k) and providing a more descriptive analysis of the characteristics of the generated segments would significantly enhance the interpretability and utility of the results for retail practitioners. Finally, the abstract effectively identifies future avenues in hybrid algorithms and additional metrics, underscoring a forward-looking perspective for advancing customer segmentation techniques.

Full Text

Comments

You need to be logged in to post a comment.