We use new data on transportation to determine its effect on the international trade of eastern and southern Africa (ESA). The effects were estimated with gravity models using alternative techniques to check for robustness. Freight forwarders provided information on freight charges (price), average duration of shipment (time) and variability in time. The new transportation variables and the more commonly used distance variable appear to have smaller effects overall than some previous research indicated, although our distance coefficient is identical to that of Coulibaly and Fotagné for West Africa. In contrast to other research critical of distance as a proxy for transport impediments, we find similar effects from distance and our measure of transport time, but both of these variables have significantly larger coefficients than price and variability. Landlocked ESA countries are twice as dependent on trade as the whole region after accounting for control variables, but they are not significantly more affected by changes in price and time, and they are less affected by distance and variability. The relative size of the time coefficients suggests that measures to lessen delays in shipping, for instance, faster border clearance, might encourage more trade than cost cutting measures that lowered price proportionately.