Global comparison services facilitate easy comparison of product offerings around the world. To offer such services, one has to address the semantic heterogeneity problems that often arise when data is collected from sources around the world. In this chapter, the authors use examples to illustrate three types of semantic heterogeneity problems that a global comparison service may encounter. Then they present a mediation architecture as a solution to addressing these problems. The feasibility of using the architecture to enable global comparison is demonstrated with a prototype application. An evaluation of the solution shows that it is scalable due to its capability of automatically generating necessary conversions from a small set of predefined ones.