The development of digital technology has brought significant changes to the e-commerce industry, including the improvement of data-driven services such as product recommendation systems. Marketplace Buket, which provides various types of bouquets online, faces the challenge of understanding consumer preferences to increase customer satisfaction and loyalty. To address this challenge, this study implements the Collaborative Filtering (CF) method as an approach in building a recommendation system capable of analyzing user behavior patterns based on consumer rating data for products. The CF method allows the system to identify user preferences by comparing similarities between consumer behaviors. By utilizing rating data, the system can recommend products that are relevant and in line with the user's interests, even if the user has never viewed or purchased the product before. This study tests the effectiveness of the recommendation system using real-world data and observing the results of recommendations given to specific users. The results show that the system can provide product recommendations with a high level of relevance, such as products I24 and I26 which are at the top of the recommendation list. In addition to providing relevant results, the system is designed with a simple interface for easy use by general users. The findings of this study indicate that the implementation of CF not only improves the quality of the user experience but also contributes to service efficiency and potential sales increases, including in intercity areas. Overall, this research provides an important contribution to the development of data-driven marketing strategies and lays the foundation for the future development of more complex recommendation systems.