Dynamic Sentiment Patterns in Social Media Crisis Communication: A Comparative Study of Starbucks and Tesla
Abstract
Nowadays, with the popularity of social media, information shifted from the previous crisis dissemination method to a more complex and rapid online spread. Public sentiment had a greater impact on shaping corporate reputation and market results. This study examines how social media emotions spread on the Internet and influence public perception during corporate crises and conducts a comparative analysis using two different cases. Due to data access restrictions, this article presents an analysis of manually collected tweet samples: 52 tweets from Starbucks' 2024 union movement and 30 tweets from Tesla's November 2023 vehicle recall. Manually mark the emotions as positive, neutral or negative based on the attitude expressed in the tweet to the company and events. The survey results reveal significant differences: Starbucks' tweets are mainly positive (56%), with strong public support for its workers, while Tesla's discourse on product safety issues is polarized (37% positive, 40% negative). These patterns indicate that labor disputes generate more persistent sympathy, while discussions triggered by product safety crises are more objective and balanced. This research contributes to crisis communication literature by emphasizing the importance of conducting specific analyses based on different types of crises and provides practical significance for enterprises' crisis response strategies and investors' decision-making. The limitations of this paper include a small sample size and the lack of quantitative analysis, which points out the direction for future research on larger datasets.
References
- Samet, U. (2024) Online reputation as an economic asset: A meta-analysis of digital perception's impact on organizational value. SSRN, 5347844.
- Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D. et al. (2019) RoBERTa: A robustly optimized BERT pretraining approach..
- Shu, X. (2024) BERT and RoBERTa for sarcasm detection: Optimizing performance through advanced fine-tuning. Applied and Computational Engineering, 97, 1-11.
- Todd, A. (2025) Financial sentiment beyond text: A multimodal approach to understanding financial market dynamics and investor behaviours. Doctoral dissertation, University of Strathclyde.
- Bartov, E., Faurel, L. and Mohanram, P.S. (2018) Can Twitter help predict firm-level earnings and stock returns? The Accounting Review.
- Cunningham, S. (2021) Causal Inference: The Mixtape. Yale University Press.
- Cartea, Á., Drissi, F. and Monga, M. (2024) Decentralized finance and automated market making: Predictable loss and optimal liquidity provision.
- Tzimiris, S., Nikiforos, S., Nikiforos, M.N., Mouratidis, D. and Kermanidis, K. (2025) A comparative evaluation of transformer-based language models for topic-based sentiment analysis. Electronics, 14(3), Article 567.
- Zhao, J., Wu, J., Feng, X., Xiong, H. and Xu, K. (2012) Information propagation in online social networks: a tie-strength perspective. Knowledge and Information Systems, 32(3), 589-608.
- Yu, W., Li, C. and Wang, T. (2025) The Impact of Regulation Threshold Heterogeneity on the Spread of Negative Emotions in Social Networks. IEEE Access.