Product recommendations on e-commerce platforms play a crucial role in supporting customers' purchasing decisions by leveraging user data to provide relevant product suggestions. With the increasing volume of e-commerce data, recommendation methods are needed that are not only accurate but also capable of being applied to diverse datasets. This research focuses on evaluating three product recommendation methods, namely User-Based Collaborative Filtering, Item-Based Collaborative Filtering, and Content-Based Filtering, using various datasets from the Kaggle platform, including transaction data and user reviews. The main problem identified is how to ensure that these three recommendation methods remain optimal despite using different datasets. Through an experimental approach, this research aims to implement and evaluate the performance of these recommendation methods. The results of this study are expected to demonstrate that one of the recommendation methods can work generally on various datasets, thereby making a significant contribution to the selection of the appropriate product recommendation method on e-commerce platforms.