We study the structure of an optimal management for innovative activities. The top management of a buyer hires a supplier for production. The production efficiency can be enhanced by investing in R&D before production. The buyer chooses between using its own subunit for the R&D, or outsourcing the task to the supplier (integration of R&D and production). Our analysis reveals that (i ) when the R&D cost is small, the buyer prefers in- house R&D, (ii ) when the R&D cost is intermediate, the buyer prefers outsourcing R&D, and (iii ) when the R&D cost is large, the buyer prefers in-house or outsourcing R&D, depending on the parameter. Within the regime in which R&D outsourcing prevails, the optimal production levels may have "partial bunching" for successful yet less favorable R&D results, depending on the R&D cost. Thus, motivating R&D may require that less favorable R&D results be overly and equally appreciated, unless the R&D is a failure.