In an increasingly digital world, the accumulation of digital files has become a growing phenomenon, with significant psychological, social, and organizational implications. This study addresses digital hoarding (DH), an emerging behavior that has traditionally been treated as a binary state: one either is or is not a digital hoarder. Using a sample of 1.490 higher education participants and applying unsupervised machine learning techniques (K-means), four clusters were identified, representing distinct types of digital hoarders, ranging from minimalist profiles to those with problematic digital hoarding. These clusters differ in key dimensions such as difficulty discarding (DD), clutter (CL), acquisition (AD), loss of control (LC), and negative consequences (NC). The results challenge the previous binary notion and suggest that digital hoarding is a complex and multifaceted phenomenon that operates on a continuum. From a theoretical perspective, this work proposes understanding digital hoarding as a continuum, opening new lines of research to explore the factors influencing the transition between different clusters. From a practical standpoint, the findings provide tools for designing organizational and technological strategies to mitigate the risks associated with digital hoarding. This work contributes to the literature on human-computer interaction (HCI) and information systems, opening new research avenues and promoting more effective interventions for digital content management.