Maximizing Social Media User Engagement Through Predictive Analytics in Retail Tourism: Identifying Key Performance Indicators That Trigger User Interactions

This study examines and evaluates key performance indicators (KPIs) that impact user engagement on social media platforms, with a primary focus on fashion retail within seasonal tourism contexts. The primary objective is to determine which engagement metrics most accurately predict user interaction levels and to enhance strategic decision-making in digital marketing. Using a dataset of 2500 Facebook photos and videos from a women’s retail store, collected between 2016 and 2024, the study employs descriptive analysis and predictive modeling. Three KPIs—such as 3 s video views, reach from organic posts, and other clicks—are examined for their impact on user engagement. The posts are categorized into engagement levels, and classification models, including Random Forests (RF), Extreme Gradient Boosting (XGBoost), K-Nearest Neighbors (KNN), and Naïve Bayes (NB), are evaluated. Results show that short video views and post reach are key predictors of user engagement. With XGBoost achieving a classification accuracy of 94.73%, the models perform effectively, and Cronbach’s alpha analysis confirms the consistency among the variables selected. The findings underscore the significance of KPI analysis in social media strategy and illustrate the value of data mining techniques in uncovering user behavior patterns that offer practical insights for optimizing digital marketing efforts.