Narrative character networks are associated with storytelling success

Narratives are central to human communication and entertainment, yet why some stories captivate audiences better than others remains poorly understood. Building on theories emphasizing character interdependence, this study introduces a network-based framework for examining storytelling success. We conceptualize narratives as networks composed of characters and their interactions and apply large-scale computational analysis to over 10,000 novels and 1,000 film screenplays. Structural features of these networks are related to storytelling success: lower clustering is associated with greater novel popularity and higher film ratings, whereas higher modularity is associated only with novel popularity and shorter path lengths only with film ratings. These findings demonstrate that storytelling success is related to how characters are structurally connected and organized. By formalizing story architecture as a measurable network, we provide a scalable framework for advancing media entertainment theory, linking cognitive models of graph learning with the structural organization of stories at scale.