Application of machine learning to the long-period analysis of multiplex internationalization strategies

This study offers a multidimensional approach to analyze the evolution of firms’ internationalization strategies throughout their lifetime. Conceptually, we contribute a new framework of multiplex internationalization to explain how firms’ internationalization choices interrelate across the three strategic dimensions of space, mode, and time. Methodologically, this study applies AI technology to reconstruct the dynamic internationalization portfolios of firms, in order to trace the continuous evolution of their spatial, entry mode, and timing strategies throughout their lifetime. We offer a long-period data pipeline to operationalize longitudinal strategic metrics, using digitized archival data on British overseas banks covering a 200-year period. The created strategic matrices visualize the changes in multiplex strategies that firms pursue through the parallel expansion of FDI and inter-firm networksto build spatially diversified multi-relational ecosystems, challenging the prevailing assumption of unidimensional internationalization strategies. Our multidimensional clustering analysis uncovered that firms’ internationalization choices overlap across mode, space, and time, thus confirming firms’ use of strategy ‘blending’ for international expansion, contrary to the dominant view of firm internationalization as a uniform process. Through ML-algorithmic modelling, we identified six strategic clusters of banks that pursued significantly different internationalization strategies as they evolved their polycentric networks through complementary entry modes, whilst employing varying timing strategies, in order to transition across diversified spatial positions. Finally, we suggest directions for further exploration and conceptualization of multiplex internationalization strategies, generalizable for firms across other sectors.