<![CDATA[The rapid growth of digital advertising platforms has encouraged businesses to adopt data-driven strategies in order to enhance the effectiveness of their promotional activities. One of the most widely used digital advertising services today is Meta Ads, which provides various performance metrics related to audience interactions. This study aims to segment Meta Ads audiences based on the effectiveness of promotional content using the K-Medoids clustering algorithm, which is known for its robustness in handling outliers compared to other clustering methods. The dataset used in this research consists of advertising access data obtained from ARTECH – PT. Arij Teknologi Inovasi. The data were processed and analyzed using RapidMiner as a data mining tool. After undergoing data preprocessing stages, including data cleaning and normalization, a total of 495 Meta Ads records were deemed suitable for clustering analysis. The results of the study show that the K-Medoids algorithm successfully grouped the data into two distinct clusters. Cluster 1 consists of 465 items and represents the dominant audience segment with relatively homogeneous interaction behavior, indicating consistent engagement patterns with promotional content. Meanwhile, Cluster 0 contains 30 items, representing a smaller but more specific audience segment with different access and interaction patterns. These findings demonstrate that the K-Medoids algorithm is effective in identifying meaningful audience segments from digital advertising data. The resulting clusters can be utilized to support more targeted digital marketing strategies, improve promotional content design, and optimize advertising budget allocation to achieve better campaign performance.]]>
Copyrights © 2026
