Optimizing Customer Retention in Urban Laundry Services: A Comparative Analysis of Machine Learning Algorithms in Jakarta's MSME Sector

In the rapidly urbanizing context of major cities such as Jakarta, the demand for efficient and rapid laundry services poses a considerable challenge, particularly within residential settings like boarding houses or dormitories in Central Jakarta, where access to communal laundry facilities is markedly constrained. Against this backdrop, laundry enterprises categorized under Micro, Small, and Medium Enterprises (MSMEs) in the service sector have witnessed significant growth. These services typically encompass the collection, washing, drying, folding, and delivery of laundry to the customer's domicile. Nevertheless, the burgeoning proliferation of laundry businesses has precipitated intensified competition within the sector, compelling these enterprises to devise strategies for customer retention and revenue augmentation. This research endeavors to delineate effective strategies and solutions to navigate the competitive landscape of the laundry industry. The study undertakes a meticulous evaluation of various machine learning algorithms, including Decision Tree, Logistic Regression, Naïve Bayes, Support Vector Machine, Random Forest, XGBoost, LightGBM, and CatBoost, based on their performance with unseen data. This evaluation employs a comprehensive set of metrics such as precision, recall, Fl-score, and ROC AUC to determine the efficacy of each model. The findings reveal that the CatBoost Classifier algorithm outperforms others, achieving a training accuracy of 95.9% and a testing accuracy of 85.7^%. The novelty of this study lies in its comprehensive comparison of multiple machine learning algorithms in the context of MSMEs in Jakarta's laundry sector, highlighting the superior performance of the CatBoost algorithm. Furthermore, the study contributes by demonstrating the practical application of these algorithms to improve customer retention strategies, enhance service quality, and optimize marketing tactics. Based on these results, the study recommends the implementation of the CatBoost Classifier model as a foundational tool for developing refined customer retention strategies, enhancing service quality, and optimizing marketing tactics within the laundry service industry.