Since Arrow (1962), spillovers from pioneer to follower in non-excludable innovations are central to our understanding of endogenous economic growth. Nonetheless, evidence of these spillovers in less-developed economies has been elusive. Our paper contributes by showing novel facts consistent with externalities in new export products. To avoid biases towards ex-post successes, we use data on the universe of customs transactions from Chile (1990Chile ( -2006. We nd that, rst, follower rms are more likely to enter a product if the pioneer rm survives exporting. More importantly, we also nd that pioneers enter and remain smaller than followers, which is indicative that the rst exporter may not be the rm that bene ts the most from the discovery. This fact is inconsistent with the currently standard view in international trade, in which the largest rm would be the rst willing to pay a homogeneous sunk cost of exporting. In contrast, our facts are consistent with the view that smaller pioneer exporters are data producers , whose spillovers bene t larger followers. We o er a simple model to formalize this intuition, based on the idea that large exporters have more choices on how to allocate their managerial capacity. This real option makes large exporters wait, as to assign their marginal manager on the best possible project. In contrast, smaller and more focused rms prefer to be pioneers. JEL classi cation: L26 ; F14; O4.