Cross cultural analysis of emotions on social media branding communication with evidence from big data

Tanaltay, Altuğ (2023) Cross cultural analysis of emotions on social media branding communication with evidence from big data. [Thesis]

[thumbnail of 10353984.pdf] PDF

Download (8MB)


Localization efforts by global brands are frequently reflected in their social media strategies, with differing degrees of adaptation on the tone, emotion, and symbols employed. Moreover, emotions play a vital role in the consumption experience, influencing word-of-mouth communications and brand loyalty. The similarities and variations in the usage of emotional content by parent brands and their local counterparts and understanding their effect on branding communication constitute an emerging field of research that can uncover the valuable insights for better business. Besides quantifying cross-cultural differences in the preferences and use of emojis as nonverbal cues regarding brand communication context among two communities at different levels of social collectivism: English and Turkish speaking Twitter users, the aim of this research is twofold. First, a language, context and time aware framework for automatically extracting basic emotions of happiness, surprise, sadness, anger, fear and disgust from multinational brands’ social media messages is proposed along with the evaluation of statistical and state-of-the-art deep learning methodologies on emotion classification domain. Second, the impact of emotions, as well as the most prominent structural and textual features on the popularity of brands’ social media messages in a cross-cultural manner is addressed using an extensive data sample of Twitter messages collected for 33 brands from FMCG, Fast Food, Technology, Automotive, Apparel, Retail, Finance, and Logistics industries over a time window of three years. In order to mitigate the issues associated with the violation of OLS iv regression assumptions, a random subspaces regression approach utilizing bootstrap resampling, which enables the quantification of feature importance and the assessment of their effects. Contributing to marketing, machine learning, social media marketing, and emotion literature, the findings are valuable for internationalizing and born-global firms to engage emotionally with their global and emerging market consumers in their social media marketing campaigns.
Item Type: Thesis
Uncontrolled Keywords: Emotion classification. -- consumer emotion. -- social media marketing. -- Twitter. -- brand post popularity. -- big data. -- global. -- local. -- emerging market. -- Duygu sınıflandırması. -- tüketici duygusu. -- sosyal medya pazarlaması. -- marka gönderi popülerliği. -- büyük veri. -- küresel. -- yerel. -- gelişmekte olan pazar.
Subjects: H Social Sciences > HD Industries. Land use. Labor > HD0030.2 Electronic data processing. Information technology
Divisions: Sabancı Business School > Management and Strategy
Sabancı Business School
Depositing User: Dila Günay
Date Deposited: 28 Sep 2023 11:38
Last Modified: 28 Sep 2023 11:38

Actions (login required)

View Item
View Item