Algorithmic Unpredictability on Social Media Platforms: A Multimethod SEM Study of Its Effects and Mitigation Strategies for Exporting SMEs

Although social media platforms have become essential for exporting small and medium-sized enterprise (SMEs) seeking to reach international customers, reliance on such platforms creates vulnerabilities. This study examines how platform power translates into platform dependence and how firms can mitigate this effect. By integrating resource dependence theory with digital platform governance, we identify algorithmic unpredictability as a mechanism linking platform power to platform dependence. Drawing on signaling theory, we propose product competitive advantage as a boundary condition that weakens the effect of algorithmic unpredictability on platform dependence. We test a moderated mediation model on Italian exporting SMEs using a multimethod structural equation modeling (SEM) approach combining partial least squares SEM and covariance-based SEM. Findings reveal that algorithmic unpredictability fully mediates the power–dependence relationship, while product competitive advantage has a negative interaction effect and a negative direct effect. Our results provide actionable insights for managers to mitigate the “platform trap.”