Despite the relevance of thermophotovoltaic (TPV) conversion to many emerging energy technologies, identifying which aspects of current TPV designs are favorable and where opportunities for improvement remain is challenging because of the experimental variability in TPV literature, including emitter and cell temperatures, cavity geometry, and system scale. This review examines several decades of experimental TPV literature and makes meaningful comparisons across TPV reports by comparing each energy-conversion step to its respective, experiment-specific thermodynamic limit. We find that peak reported efficiencies are nearing 50% of their thermodynamic limit. Emitter-cell pairs that best manage the broad spectrum of thermal radiation exhibit the best efficiencies. Large gains in peak efficiency are expected from further suppression of sub-bandgap radiative transfer, as well as improvements in carrier management that address bandgap underutilization and Ohmic losses. Furthermore, there is a noticeable practical gap between the leading material pairs and integrated devices, mainly due to a lack of scaled-up high-performance materials, which exposes surfaces to parasitic heat loss. Provided these challenges are overcome, TPVs may ultimately provide power on demand and near the point of use, enabling greater integration of intermittent renewables.