Adaptive Behavioral Nudging Under Information Overload: A Context-Aware Digital Marketing Framework

Consumers on digital platforms often choose under information overload, yet most behavioral nudging systems still rely on static preference profiles that ignore how cognitive load, time pressure, and distraction alter choices within a session. This study proposes the Context-Anchored Temporal Preference Inference (C-TPI) framework, which combines evidence-accumulation accounts of preference formation with context-aware computing. The framework uses a context-gating mechanism to adjust preference estimates to situational signals and a mixture-of-experts layer to separate decision regimes, such as high- versus low-load browsing. It was evaluated on three consumer behavioral datasets (e-commerce transactions, mobile application sessions, and nudge A/B tests) against four baseline models. C-TPI reached a Precision@5 of 0.487 and a Contextual Alignment Score of 0.763, compared with 0.453 and 0.715 for the strongest baseline, and its accuracy declined by 9% from the lowest to the highest information-load quartile, against 31% for a recurrent baseline. Ablation results indicate that context gating accounts for the largest share of these gains. The findings support treating preferences as context-dependent temporal processes and suggest matching nudge complexity to cognitive load. Because the same choice-architecture logic is increasingly applied inside organizations, the study also discusses implications for electronic human resource management (e-HRM), including benefits enrollment, learning recommendations, and the protection of employee data. Limitations concern data privacy, algorithmic transparency, and cross-cultural generalizability.