Knowledge sharing is a difficult task for most organizations, and there are many reasons for this. In this article, we propose that the nature of the knowledge shared and an individual's social network influence employees to find more value in person‐to‐person knowledge sharing, which could lead them to bypass the codified knowledge provided by a knowledge management system (KMS). We surveyed employees of a workman's compensation board in Canada and used social network analysis and hierarchical linear modeling to analyze the data. The results show that knowledge complexity and knowledge teachability increased the likelihood of finding value in person‐to‐person knowledge transfer, but knowledge observability did not. Contrary to expectations, whether the knowledge was available in the KMS had no impact on the value of person‐to‐person knowledge transfer. In terms of the social network, individuals with larger networks tended to perceive more value in the person‐to‐person transfer of knowledge than those with smaller networks.