Semantic analyses vs. IPC co-classification analyses of patents: Which one better serves to anticipate converging industries?

摘要

The convergence of industries exposes the involved firms to various challenges - e.g., changes in technological paradigms or even in entire value chains. In such a setting, a firm's response time becomes key for its future success. Hence, different approaches to anticipate convergence have been developed in the recent past. Building on the well-accepted notion that - at least in research intense industries - technology convergence precedes industry convergence, patents play an important role as precursors of new technological developments. So far, particularly IPC co-classification patent analyses have been successfully applied in different industry settings. However, these analyses are only applicable on a broader industry/technology level and are of a high quantitative value, but do not provide detailed information on the convergence process. Here, we aim at developing a concept to anticipate convergence even in small samples and simultaneously providing more detailed information on its origin and direction. To this end, we analyzed the semantic similarity of 326 US-patents on phytosterols. In line with our argument that an increasing semantic similarity over time across patents of formerly different technology fields can be taken as an indicator for convergence, we could find evidence for convergence of the food, pharmaceutical and personal care sectors.