K-Means and Decision Tree Models for Analysis Students Perception of Digital Marketing

Abstract

Amid intensifying competition among educational institutions and the rapid expansion of digital platforms, schools must adopt effective digital marketing strategies to attract prospective students. This research examines students’ perceptions of the implementation of digital marketing at the Global Islamic Boarding School (GIBS) in Barito Kuala Regency, South Kalimantan, Indonesia. Data were collected from 95 students using a structured questionnaire comprising 20 variables related to digital marketing and word-of-mouth communication. A hybrid analytical framework integrating K-Means Clustering and Decision Tree analysis was employed to identify perception-based segments and determine key differentiating indicators. The Elbow method indicated an optimal three-cluster solution, resulting in highly positive, negative, and moderate perception groups. Decision Tree results revealed that engaging content (B7), social media stability (B12), and invitation to attend school (C19) were the most influential variables distinguishing clusters, with satisfactory classification performance supporting the robustness of the segmentation. The findings suggest that emotionally engaging and experientially rich content plays a more decisive role in shaping student perceptions than purely technical aspects of digital media management. Accordingly, schools should optimize their digital marketing strategies by emphasizing authentic student experiences, maintaining consistent, stable online communication, and incorporating persuasive calls to action to enhance prospective students’ interest.