Research on negative brand relationships has gained traction over the years. However, most of this research has focussed on the cases involving product or service failure by a firm, leaving conflicts arising as a result of values and ethics espoused by a firm relatively unexplored. In this research, we focus on one such case and explore the factors that influence virality of social media content in the context of Consumer Brand Sabotage (CBS). Based on a week of Twitter data pertaining to specific hashtags associated with a brand’s sabotage, we explore how tweet related attributes affect the potential for amplification of the tweets related to the event. We categorize the factors as informational, interactional, and creator specific, and build machine learning (ML) models to predict the retweet likelihood of CBS tweets. We find that while informational factors associated with the tweets (such as, hashtags, URLs and emotions) are important to predict the diffusion of CBS-related tweets, this was not the case for interactional factors (such as, reply, like, quote, etc.). For creator factors, we found that considering the number of followers of the creator in the ML models reduced the predictability of diffusion of CBS-related tweets, and found verified accounts to be of little importance as well. We discuss the implications of these findings for practice and research, and present scope for future research.

多年来，对负面品牌关系的研究越来越受到关注。然而，大多数研究都集中在涉及公司产品或服务失败的案例上，而由于公司所奉行的价值观和道德规范而产生的冲突相对未得到探索。在这项研究中，我们重点关注这样一个案例，并探讨在消费者品牌破坏 (CBS) 背景下影响社交媒体内容病毒式传播的因素。根据与品牌破坏相关的特定主题标签的一周 Twitter 数据，我们探讨了推文相关属性如何影响与该事件相关的推文被放大的可能性。我们将这些因素分为信息性因素、交互性因素和特定于创作者的因素，并构建机器学习 (ML) 模型来预测 CBS 推文的转发可能性。我们发现，虽然与推文相关的信息因素（例如主题标签、URL 和情感）对于预测 CBS 相关推文的传播很重要，但对于交互因素（例如回复、点赞、引用、 ETC。）。对于创作者因素，我们发现在 ML 模型中考虑创作者的关注者数量会降低 CBS 相关推文传播的可预测性，并且发现经过验证的帐户也不太重要。我们讨论这些发现对实践和研究的影响，并提出未来研究的范围。