Big Data Analytics And Firm Performance: A Text Mining Approach

Big data analytics, used to be a state-of-the-art process, now it is a must. Nowadays, firms promote this as a standard operating procedure. It is not only the luxury of the large multinationals, but we see, that also local companies from the Romanian economy start to present this in their annual financial reports. Numerous studies tried to measure and prove the positive impact of big data analytics upon firm’s performance. In the current study we do not aim to do the same. Instead, we propose to review all the studies on this topic, which we could find, and to test a common practice in the field of big data analytics called sentiment analysis, which is used to assess the attitude of the writers towards their subject. Our hypothesis is that there is a strong relationship between the reported results of the studies, called effect sizes and the wording or phrasing of the articles, called sentiment score. As a methodology, we used Publish or Perish which is an academic “search engine to search all the articles on the big data analytics – firm performance topic available on Google Scholar between 2010-2020. Our initial search found over5500 articles, books and other publications. Based on our criteria, we kept 120 articles. The main finding is that there is no correlation between the effect sizes of the articles and the sentiment score we obtained. Moreover, we found that on average, based on the wording used, the studies tend to overstate their results.