Researchers have long studied the effect of network embeddedness on innovation outcomes, resulting in mixed findings. In this paper, we aim to take a novel approach to reconcile prior conflicting perspectives. We argue that previous seemingly discrepant findings may stem from ignoring different forms of network embeddedness and innovation. An innovator has doubly embedded relations in innovation networks: relational and structural embeddedness (SE). Additionally, innovations are categorized into two distinct types: exploratory and exploitative innovations. Our analysis is based on 2936 patent records involving 1561 inventors from a large U.S. company over a 16-year period. The findings indicate that, both relational and SE contribute to promoting exploitative innovation, while they have inverted U-shaped relationships with exploratory innovation. Furthermore, we argue that the whole network structure (e.g., the small-world quotient) must be considered to fully understand the effectiveness of network embeddedness. The implications of our findings are discussed for embeddedness, innovation, and network theory. For policy practice, researchers and knowledge workers should be more aware of modifying the levels of relational and SE, adopt appropriate embeddedness levels in a different situation, and match their embeddedness strategy with the whole structure of the intrafirm network.
Stored only when both papers are in the corpus — 75 citations exist upstream.
Imported from OpenAlex (CC0) — nothing here is generated.