Empirical PLS SEM Modeling of Digital Marketing Impact in Social Commerce Systems
Abstract
Social commerce systems require precise modeling of how content cues and source features drive consumer purchases. This study evaluates the structural impact of customer reviews, influencer endorsements, and influencer credibility on purchase intentions for Skintific cosmetic products in Indonesia, with brand trust examined as a mediating variable. Primary data were collected from 190 active social commerce users through an online Google Forms questionnaire using purposive sampling. Structural equation modeling was conducted using SmartPLS 4.0 based on partial least squares with a 5,000-resample bootstrap procedure. The measurement model confirmed convergent validity, with average variance extracted values ranging from 0.651 to 0.718 and composite reliability (ρc) values between 0.881 and 0.899. Heterotrait-Monotrait ratios below 0.85 established discriminant validity, while inner variance inflation factor values below 2.042 indicated the absence of collinearity. The empirical results showed that customer reviews, influencer endorsements, and influencer credibility significantly and positively influenced trust, with path coefficients of 0.251, 0.381, and 0.251, respectively. Trust also demonstrated a strong positive effect on purchase intention with a path coefficient of 0.640. The mediation analysis confirmed that trust mediates the relationships between customer reviews, influencer endorsements, and influencer credibility with purchase intention, producing indirect effect coefficients of 0.160, 0.161, and 0.244, respectively. These findings suggest that digital commerce strategies should position influencer endorsements as trust-building mechanisms rather than merely promotional activities.