Exploration and Exploitation in Dynamic Knowledge Transfer Networks
Cesar Garcia-Diaz, Mohsen Jafari SonghoriOrganizational performance depends on internal processes of knowledge generation, transfer and coordination (Bae &Koo, 2008). Not only does knowledge provide a foundation of the firm (Grant,1996), but also its successful transfer is considered a key determinant of organizational performance (Levine & Prietula, 2012). Successful knowledge transfer depends on structural characteristics of the organization (Reagans & McEvily, 2003; Nickerson & Zenger, 2004). Structures also have a dynamic nature since individuals’ actions simultaneously shape and are shaped by such structures, implying that knowledge transfer and structure jointly coevolve (Luo et al., 2015). This work is aligned to current research efforts that aim to understand co-evolutionary dynamics between individuals’ actions and network structure.
We investigate the bidirectional effects of individuals’ decisions on the development of network structure as well as the effects of such a structure on the (imperfectly rational) decisions–in terms of knowledge transfer–made by network members. Our approach takes into account the fact that organizational knowledge is generated by the individual (Davis & Aggarwal, 2019). To investigate the effectiveness of organizational knowledge transfer, we focus on how a fitness landscape search is performed by heterogeneous computational agents that, in a distributed fashion, share and assimilate information about the landscape, while simultaneously adjusting their local structure.
Our work departs from previous perspectives of knowledge transfer in the literature in at least three aspects:
(i) The extant literature usually conceives knowledge transfer as an information transfer process, even if this transfer is constrained by absorptive capacities (e.g. Davis & Aggarwal, 2019). We incorporate two dimensions of knowledge transfer: the traditional informational dimension, and the cognitive dimension (Knudsen & Srikanth, 2014), indicating that epistemic interdependence (Puranam et al., 2012) is a key element in knowledge transfer processes.
(ii) The approach is dynamic. Although some works have highlighted the importance of the interplay of knowledge and organizational structure dynamics, they do not explicitly incorporate individual-level rules of behavior. Traditional research in knowledge transfer has only considered static network structures, with some few exceptions (Chang & Harrington, 2007; Luo et al., 2015).
(iii) Considerations of how individuals search and use knowledge within the organization, which we call intra-group exploitative and exploratory behavior, depend on how individuals adjust their local social structure. Exploration and exploitation are usually considered as organizational-level characteristics, but not as individual-level strategies. Here, we do not only explore what individual-level exploration and exploitation imply for collective performance, but also we investigate how individual-level exploration and exploitation add up to organizational-level exploration and exploitation (the aggregation problem). This approach complements works done at the firm level, and the efforts for studying knowledge mobilization from the individual to the organizational level. Our work links micro and macro levels of organizations, and complements the organization learning literature by revealing the implications of collective knowledge adoption in connection to individual-level actions.
We provide insights on (i) the determinants of successful intra-firm knowledge transfer, (ii) the causal link between the collective and the individual exploration-exploitation balance, and (iii) the connection between intra-group knowledge and structure dynamics.