Self-repair Mechanism of User-Generated Content (UGC) in Brand Crises: A Computational Communication Model Based on BERTopic and Survival Analysis

In today's digital landscape, where social media plays a pivotal role, user-generated content (UGC) has emerged as a crucial element in the management of brand crises. This research investigates the mechanisms through which UGC can self-repair during such crises, utilizing BERTopic for topic modeling alongside survival analysis to assess both the durability and efficacy of UGC. Our findings reveal five predominant themes in crisis-related UGC: product quality, customer service, brand trust, crisis response, and social impact, which together account for a significant 78.3%of the overall content volume. The survival analysis indicates that certain characteristics of content, notably originality and the inclusion of multimedia elements, substantially bolster the persistence of UGC; specifically, original content is associated with a hazard ratio of 0.64 (p < 0.001), while multimedia content has a hazard ratio of 0.71 (p < 0.001). Furthermore, structural equation modeling illustrates that the self-repair mechanisms of UGC—such as consumer advocacy, narrative correction, and sentiment recovery—are vital for brand recovery, with consumer advocacy proving especially effective in contexts related to performance issues (β = 0.45). The insights derived from this study challenge conventional crisis management approaches by emphasizing the necessity of nurturing organic consumer narratives and harnessing the innate recuperative capabilities of UGC ecosystems. Ultimately, this research enriches the field of brand crisis management by offering a detailed framework for understanding UGC dynamics and providing practical recommendations for formulating more effective crisis response strategies.