The performance of many photoactive devices is impacted by the dynamics of excitons within the photoactive layer. Typically, these dynamics are deduced by independently conducted transient absorption measurements on films outside of the device with the assumption that the photophysics in free standing films inform the photophysics of the film inside the device. But there are interfaces in devices that are not present in films and the morphology and atomic packing might be impact by device construction. Moreover, photoabsorption experiments measure all excitons whereas the best photovoltaic devices are perhaps 20% efficient, meaning that only 1 in 5 excitons actually produce electricity. Thus, the conclusions drawn from free standing films might not be directly relevant to devices. This talk will cover results made possible by new advances in ultrafast 2D photocurrent spectroscopy that enable ultrafast exciton dynamics to be directly compared between films and devices. In photovoltaics made from semiconducting carbon nanotubes, we find that the conclusions drawn from films versus devices are quite different, leading us to reinterpret long standing experiments and reevaluate our goals for device architecture.