Learning ambidexterity and technology innovation: The moderating effect of knowledge network modularity

Abstract

The aim of this paper is to investigate the impact of learning ambidexterity on firms’ technological innovation, especially in the context of varying degrees of modularity in firms’ knowledge networks. By combining temporal and cognitive aspects, this study focuses on the micro-level utilization of knowledge and redefines exploratory and exploitative learning as learning unfamiliar and familiar knowledge, respectively. Building upon previous research, we classify learning ambidexterity into combined and balanced forms, further investigate their influence on innovation performance. Utilizing a panel data set spanning 10 years of 63 semiconductor firms listed in the US market, our findings reveal that combined learning ambidexterity effectively harnesses the complementary effects of both learning patterns, resulting in a positive influence on technological innovation. Conversely, balanced learning ambidexterity hampers knowledge transferability between these two patterns leading to a negative impact on technological innovation. Additionally, we employ patent data to measure the degree of modularity within firms’ knowledge networks. Our results indicate that higher levels of modularity enhance the positive effect of combined learning ambidexterity on innovation while mitigating its negative impact for balanced learning ambidexterity. These findings suggest the importance for firms to strategically manage and dynamically orchestrate learning ambidexterity alongside double-loop learning practices, while continuously structuring and facilitating reuse within organizational knowledge networks to create favorable circumstances for the effective implementation of learning ambidexterity.