Keywords
How to Cite
Abstract
The article examines order cancellations in the construction tools segment (Power Tools and Hand Tools) across major marketplaces during the period from October 1, 2024 to March 31, 2025. The study is based on real data from an online store, including daily orders with several transactions recorded for each marketplace.The main focus is on cancellations that happen before the customer receives the product or at the pickup point. These cases are seen as early signs of issues in logistics and service quality. To analyze the factors that influence cancellations, machine learning methods were used, such as logistic regression, Random Forest, and gradient boosting. The models were compared using classification accuracy and ROC-AUC metrics. The results show that cancellation rates differ between marketplaces. They also confirm that it is possible to build a model that can predict cancellations in advance. This can help reduce operational costs and improve sales management. Overall, the practical value of the study is in developing a tool for predicting cancellations and increasing efficiency in managing sales within the construction tools segment.