The advancement of technology drives online shopping activities in society. There are 59 million SMEs in Indonesia, 56.3% of which have utilized social media platforms such as Instagram, Facebook, and TikTok. According to a TikTok Indonesia survey in 2024, 70% of TikTok users have watched live shopping, and less than 40% have engaged in impulse buying. This study aims to analyze the effect of impulse buying on the sales growth of SMEs in Indonesia while enriching social commerce literature by integrating a machine learning approach to predict impulse buying and comparing the performance of classification algorithms in the context of consumer behavior. Data were obtained through simulations of 1,000 TikTok Live Shopping SME sessions from January to March 2025 across various product categories. Next, an analysis of machine learning methods was conducted using the Support Vector Machine (SVM) Algorithm, Naïve Bayes Classifier, and K-Nearest Neighbor, as well as calculations with simple linear regression to test the predicted impulse buying and sales growth. The SVM results had the highest accuracy rate of 91.2%, and linear regression showed a significant positive impact between impulse buying and MSME sales growth (β = 0.64; R² = 0.72; p < 0.001).
Copyrights © 2025