Batteries based on the cell reaction between alkali metals and oxygen are highly attractive for energy storage due to their superior theoretical energy density. However, despite continuous progress, fundamental challenges in the further development of these cell systems remain. Understanding the oxygen electrode reaction and improving cycle life, while at the same time maximizing the practical energy density, are some of the most important issues that need to be addressed. Here, the product formation in aprotic sodium-oxygen cells is studied and it is shown how cycle life and practical capacities can be improved. Different cell reactions (leading to either NaO2 or Na2O2 as discharge products) have recently been reported. To understand whether the carbon structure or the local current density has any influence on the product stoichiometry or the cell performance, several carbon materials with a broad range in properties are tested. Phase-pure NaO2 is always found as a discharge product, but capacities range from 300 to values as high as 4000 mAh g(C)−1 depending on the type of carbon. More importantly, the cycle life of Na/O2 cells can be largely improved by shallow cycling, steadily yielding capacities of 1666 mAh g(C)−1 for at least 60 cycles using a Ketjen black carbon electrode.