Design/methodology/approach: Thirty one respondents from the public and private sector were interviewed using a qualitative approach. Data were analyzed using the Atlas ti qualitative software version 7.5.18 to generate themes for information sources, types and channels. Social network analysis was used to measure the actor centrality positions and influence in the IP network.

Findings: Results revealed seven main categories of actors in the Coffee IPs who shared information on coffee inputs, agronomic practices, processing and markets through three main channels. Level of cohesion was less than 10% which had negative implications on the knowledge flow, trust and collaboration among the actors. Influential positions were occupied by the processors and farmer leaders in IPs in the southern districts of Luwero and Rakai, while nursery operators were most influential in IPs of the western districts of Ntungamo and Bushenyi. Weak linkages within the social networks indicated that initiatives of the actors were fragmented, as each actor acted as an individual detached from the platform activities limiting inter-actor knowledge sharing.

Practical implications: Innovation intermediaries should focus on integrated systemic and innovative approaches to strengthen actor social linkages for knowledge sharing and better platform performance.

Theoretical implications: Actor positions and relationships in innovation networks are critical tenets for fostering knowledge exchange and performance. In an innovation platform, diverse actors are multiple sources for accessing information within a given social and institutional context.

Originality/value: The study contributes to existing debate and knowledge on institutional change in agricultural innovation systems.  相似文献