We develop a dynamic model of international business-to-business transactions in which sellers and buyers search for each other, with the probability of a match depending on both individual and aggregate search effort. Fit to customs records on U.S. apparel imports, the model captures key cross-sectional and dynamic features of international buyer-seller relationships. We use the model to make several quantitative inferences. First, we calculate the search costs borne by heterogeneous importers and exporters. Second, we provide a structural interpretation for the life cycles of importers and exporters as they endogenously acquire and lose foreign business partners. Third, we pursue counterfactuals that approximate the phaseout of the Agreement on Textiles and Clothing (the “China shock”) and the IT revolution. Lower search costs can significantly improve consumer welfare, but at the expense of importer profits. On the other hand, an increase in the population of foreign exporters can congest matching to the extent of dampening or even reversing the gains consumers enjoy from access to extra varieties and more retailers.