Scholars and practitioners increasingly recognize the importance of microblogs in capturing eWord of Mouth (eWoM) and their predictive power for cryptocurrency markets. This research in progress paper examines the extent to which microblog messages are related to bitcoin fluctuation. Building on information systems and finance literature, we examine the interactions between influencers’ extreme sentiment and the bitcoin fluctuation using natural language processing techniques and hypothesis testing. Our preliminary results show when influencers express extreme sentiment, in favour or against bitcoin, it is less likely that their tweets are related to future bitcoin fluctuation. However, when their extreme tweets are in-depth and unique, this negative relationship is moderated. Overall, our findings reveal that influencers’ sentiment is an important predictor in determining bitcoin fluctuation, but not all tweets are of equal impact. This study offers new insights into social media and its role in the cryptocurrency market.