The myth of data analysis: can patent data analysis really support patent management?

Patent data is widely used by institutions, including corporations, governments, and research institutes, to present technology strategies and policies. Despite the abundance of literature on the indispensability, significance, and constraints of patent data analysis, most studies have examined only selected subsets of patent records. By contrast, the present study analyzes the complete corpus of patents filed and granted between 2001 and 2021 at the USPTO, KIPO, EPO, and JPO. This study reviewed the relevant literature to formulate and validate six hypotheses pertaining to the analytical utilization of patent data. The goal of this study is to draw inferences regarding the significance of patent information and the value of patent analysis through the statistical examination of diverse patent sources. It was found that the existence of a time lag between R&D (research and development) investment, R&D activities, and the registration of patents renders patent data potentially outdated for the purpose of investigating the current state of technology. Additionally, there exists a strong correlation between the number of patent citations and patent renewal fees. Furthermore, given the infrequency of human errors in patent classification, patent data may be considered a reliable source for patent analysis. There is evidence of the presence of bias in the distribution of patents across various industries, and certain sectors, such as ‘agriculture', ‘mining and quarrying', exhibit a significant reliance on trade secrets as a means of safeguarding technological activities. A similar pattern was observed in the distribution of patents across sectors in the databases of main patent offices. The results of this study can help researchers and practitioners relevantly analyze technology through the use of patent data.