This work addresses how the number of suppliers employed by a manufacturer relates to its use of electronic procurement (e-procurement). Using data from the U.S. manufacturing industry, we find a surprising result that there is non-correlation between e-procurement and number of suppliers at the aggregate level. However, when we distinguish the type of goods purchased, we find that e-procurement is associated with using more suppliers for custom goods while fewer suppliers for standard goods. The finding for custom goods may sound contradictive to the prior literature proposing a "move to the middle," but is consistent with a TCE-based notion that custom goods procurement involves more asset-specific relationships than standard goods. Further, the positive association between e-procurement and number of suppliers of custom goods is negatively moderated by the degree of buyer-supplier systems integration, which is consistent with the argument that deeper integration of buyer and supplier information systems can help buyers obtain better "fit" for their customized requirements. This is an alternative to achieving better fit by employing more suppliers as proposed in the extant literature. Keywords: supply chain relationships, number of suppliers, electronic procurement, transaction costs economics, custom goods, systems integration