AI-powered personalization and customer loyalty in Abuja's e-commerce sector: The competing mediating effects of perceived relevance and privacy fatigue

Abstract

AI marketing can win and repel consumers at the same time; however, little is known about how this happens in African e-commerce. This is important because Nigeria’s e-commerce market is worth over $2 billion, with over 30 percent contributed by Abuja. Understanding how personalization drives customer loyalty is critical for businesses and policymakers. This study employs a quantitative cross-sectional design with 300 e-commerce consumers in six area councils of Abuja via a structured questionnaire. Six hypotheses, including direct effects, mediating effects, and competing indirect pathways, were tested using partial least squares structural equation modeling. The study showed a positive direct effect of AI-powered personalization on customer loyalty (β = 0.31, p < 0.001). Perceived relevance demonstrated a positive indirect effect (β = 0.42, 95% CI [0.31, 0.53]), while privacy fatigue demonstrated a negative indirect effect (β = -0.19, 95% CI [-0.27, -0.11]). The positive indirect effect through perceived relevance was significantly stronger than the negative indirect effect through privacy fatigue (difference = 0.61, 95% CI [0.45, 0.77]). The model explained 58 percent of the variance in customer loyalty. The simultaneous mediating mechanisms within a unified dual mediation model extend the personalization literature. The introduction of privacy fatigue as a mediator extends Privacy Calculus Theory beyond static privacy concern measures. Practically, e-commerce managers should enhance perceived relevance while ensuring transparent data governance. Policymakers should ensure that intervention designs balance innovation with consumer protection. This study identifies perceived relevance and privacy fatigue as competing pathways linking AI-driven personalization to customer loyalty in African e-commerce.