AREC Model: Trust dan Engagement terhadap Conversion Intention pada Social Commerce

The development of algorithm driven social commerce has enhanced platforms’ ability to deliver personalized product recommendations; however, high exposure to recommendations does not always translate into transactional intention, as consumers still consider relevance, trust, and engagement. This study aims to examine the effects of algorithmic relevance and trust on conversion intention through engagement among social commerce users in Medan City. The study adopts an explanatory quantitative approach with a cross-sectional design. Data were collected through questionnaires from 222 respondents and analyzed using Structural Equation Modeling–Partial Least Squares with SmartPLS 4 through outer model evaluation, inner model evaluation, path coefficient testing, and indirect effect analysis. The novelty of this study lies of the Algorithmic Relevance Engagement Conversion Model, which integrates algorithmic relevance and trust as antecedents, engagement as a mediating variable, and conversion intention as the outcome within a single conversion mechanism. The findings indicate that algorithmic relevance and trust have positive and significant effects on both engagement and conversion intention, while engagement exerts the strongest effect on conversion intention. Furthermore, engagement significantly mediates the effects of algorithmic relevance and trust on conversion intention. These findings extend the application of Social Exchange Theory in algorithm-based social commerce and provide practical implications for platforms and businesses to improve recommendation relevance, trust, and user engagement as strategies to enhance conversion. This study contributes the AREC Model as an integrated explanation of how algorithmic relevance and trust relate to conversion intention directly and through engagement, offering social commerce platforms and businesses a basis for prioritizing relevant recommendations, strengthening trust, and designing interactions that foster user engagement.