Digital sovereignty policies are increasingly reshaping the cross-border operations of AI firms. This study examines how digital sovereignty pressure affects the nonmarket strategic adaptation of multinational AI firms. Using 310 firm-year observations from 51 AI-related listed firms from 2018 to 2024, this study uses disclosure texts, policy events, lobbying records, and financial data to construct firm-level measures and estimate panel regression models. The results show that digital sovereignty pressure significantly increases AI firms' overall nonmarket strategy intensity, with a stronger effect on compliance-oriented strategies than on influence-oriented strategies. Further analysis indicates that data intensity strengthens the effect of digital sovereignty pressure on influence-oriented nonmarket strategy, whereas internationalization receives only partial support as a moderator. The findings advance digital sovereignty research by conceptualizing it as a firm-level information governance pressure that shapes AI firms' external strategic choices.
Existing research often treats cryptocurrency adoption as uniform, overlooking differences among assets and users. This article develops a typology linking adopter profiles to institutional orientation, financial capability, technological embeddedness, and asset affinity. The article develops a conceptual typology through taxonomy-development logic, anchored in behavioral-institutional orientation: how individuals engage with cryptocurrency as a financial and socio-technical alternative. The typology identifies three ideal types: long-term investors, linked to bitcoin as a portfolio asset; resilient crypto traders, linked to stablecoins for monetary protection and resilience; and crypto-natives, linked to Ethereum/DeFi for participation, autonomy, and experimentation. The article moves beyond monolithic technology-adoption models and deconstructs the “cryptocurrency user” as uniform. It shifts focus from trust in technology to trust in the financial status quo, highlighting institutional distrust and inflation as segment-specific drivers.
The fashion industry is resource-intensive and causing pollution, waste, and resource depletion. Sustainable fashion arose to reduce these environmental and social harms. This study examines key factors that influence user satisfaction with sustainable fashion apps. Latent Dirichlet allocation topic modeling was applied to 84,188 user reviews of sustainable fashion apps on the Google Play Store. The analysis identifies ten key factors, including app performance, service experience, and shipping, that influence user satisfaction. These are linked to the Delone and McLean Information Systems success model to evaluate their impact on post-usage satisfaction through multiple regression analysis. The study presents a novel methodology that combines text mining and regression analysis, providing unique insights into user satisfaction with sustainable fashion apps. It improves sustainable fashion marketing literature and offers practical guidance for application development to boost user satisfaction.
Promoting the digital transformation of the manufacturing industry has become a focus for advancing the digital economy alongside the real economy. However, current research has not identified the key dimensions of digital capability in manufacturing enterprises, and there is a lack of well-developed measurement scales. This study addresses this gap by conducting semi-structured interviews and analyzing secondary data from 45 middle and senior managers across 13 manufacturing enterprises. Using procedural grounded theory, it explores the key dimensions of digital capability in these enterprises, which include digital perception capability, digital business capability, and digital co-creation capability. Based on these findings, the study developed a measurement scale consisting of three dimensions and 20 items, validated through exploratory factor analysis (N=268) and confirmatory factor analysis (N=345). This research provides an effective tool for measuring the digital capability of manufacturing enterprises and offers practical insights for further empirical studies in this area.
Social media data offer international businesses real-time insight into public sentiment across countries, but dataset size makes conventional clustering difficult to apply directly. This study demonstrates a scalable workflow combining sentiment analysis, self-organizing maps, and spectral clustering to identify country-level sentiment profiles in large social media datasets. Using a dataset of 63 million COVID-19-era tweets, sentiment valence and four emotion scores—anger, fear, sadness, and joy—were analyzed across six countries. The results reveal three broadly similar profiles: a dominant neutral or low-joy profile; a mid-sized negative, high-fear profile; and a smaller positive, low-negative-emotion profile. Countries differed mainly in emotional intensity within these shared profiles. The study contributes a scalable approach for profiling country-level sentiment patterns to inform international business research and practice. These profiles should be interpreted alongside cultural, institutional, economic, and market information, not as direct measures of cultural values.
