Optimal modular design policy for complex systems in considering coordination costs

There have already been a number of managerial and engineering articles that focus on modularity, but none on the issues that the optimal modularity level of complex systems poses in terms of real options and coordination theory. To fill this gap, we develop a modularity profit model based on real options and coordination theory to examine the trade-off between modularity values and costs especially coordination costs. Two key findings are that: (1) In an expensive test and integration environment, it's optimal for an organization to divide a system into fewer modules and run fewer experiments on each, and further reduce the number of independent parallel design when unit coordination cost is high; (2) In a sophisticated test and integration environment, the optimal splitting policy depends on coordination factors such as ex-ante coordination effectiveness and bug-fixing cost, but the optimal substitution policy only depends on unit design cost.