텍스트 마이닝 기법을 활용한 코로나-19 펜데믹 발생 이전과 이후 스타트업 성장요인에 관한 인식변화 연구

Abstract

The purpose of this study is to investigate and analyze the changes in social perceptions of startup growth factors in the period before and after the COVID-19 pandemic using text mining techniques, and to provide useful information for government agencies' policy formulation and companies' decision-making for startup growth in the post-COVID-19 era. The TEXTOM program was used to collect and analyze the data. The analysis methods used were word cloud analysis and frequency analysis through text mining, TF-IDF analysis, N-gram analysis, ego network analysis, and CONCOR analysis. The results of the analysis showed that the words that changed significantly after the outbreak of the COVID-19 pandemic were “finance” from 6th to 12th, “startup” from 10th to 20th, “scale-up” from 16th to 4th, and “innovation” from 17th to 11th. Keywords such as “big business”, “government”, and “economic growth” have not appeared in the top 30 keyword impression frequency since the outbreak of the COVID-19 pandemic. This indicates that the external environmental shocks brought by the COVID-19 pandemic require new ways of growth and innovation, and that startup growth has shifted from 'exogenous structural changes' such as government support to 'autogenous structural changes' based on the startup ecosystem. In addition, since the outbreak of the COVID-19 pandemic, words related to startup growth strategies such as “accelerator”, “power”, “growth platform”, “follow-up support”, “need”, “secure”, “technology”, “barrier”, “foundation”, “market creation”, and “economy” have appeared. Through the CONCOR analysis, we concluded that social interest in the growth of startups is increasing amidst the low growth of the Next Normal, and that internal resource capacity enhancement and the scale-up ecosystem, an environmental factor that can support it, and horizontal networks among firms are key factors as the main drivers of startup growth, and discussed theoretical, policy, and practical implications based on this. Finally, we discuss the limitations and future directions of this study.