Electronic word-of-mouth (eWOM) significantly influences consumer decision-making because prospective viewers rely on peers’ first-hand opinions to reduce uncertainty and anticipate their own enjoyment—an effect that is especially pronounced for narrative prod- ucts such as movies. This influence is amplified by Web 2.0 technologies, which make peer reviews highly searchable at critical decision points, and by the credibility conferred through identity-based social networks. However, despite this growing social reliance, most existing studies overlook the broader context in which reviews circulate, implicitly treating each review as an independent unit. This study investigates how content features, reviewer characteristics, social network position, spoiler labels, and herd behavior collectively shape the perceived usefulness of movie eWOM on Douban Movies. By analyzing data from 185,516 movie reviews and their associated user networks using zero-inflated negative bi- nomial regression models, this research demonstrates that content richness and reviewer credibility substantially increase perceived usefulness. Reviews containing spoilers receive higher usefulness evaluations, suggesting spoilers reduce consumer uncertainty. Notably, re- viewers occupying bridging positions (high betweenness centrality) exert greater influence, aligning with social capital theory. In contrast, content similarity driven by herd behavior negatively affects usefulness perceptions, indicating redundancy diminishes perceived infor- mational value. This research extends the eWOM literature by integrating content-level and individual-level characteristics with interaction dynamics, offering practical implications for managing online review platforms and narrative product marketing strategies.