Utilizing a global IT sourcing strategy bears enormous growth potential. With the main focus on cost reduction in valuation of sourcing alternatives, risk and risk diversification effects are often inadequately considered or completely neglected. This systematically results in wrong decisions about global sourcing. Correct decisions are in particular important for the success of IT services providers (ITSP), which are the major beneficiaries of the market growth, though being faced with intensifying competition. This paper proposes a decision model for allocating software development projects of an ITSP to available sites in a risk/cost efficient way by adapting Markowitz’s Modern Portfolio Theory to IT sourcing decision making. The suggested approach covers not only costs and sourcing risks but also interdependencies between both sites and projects. Additionally, we propose methods for quantifying the necessary input parameters. We demonstrate the practicability of our approach in a case study with data from a major ITSP.