Standard unfocused Synthetic Aperture Radar (UFSAR) nadir-altimeter products with a posting rate of 20 Hz provide water level heights with an along-track spacing of 300 meters, which is far from being optimal for coastal zone and inland studies. The fully focused FFSAR processing, which allows much higher resolution and posting rate, shall resolve small scales interferences along-track and thus improve the accuracy and precision of the geophysical estimates. This study looks at different water surfaces in rivers, lakes and coastal zones to investigate under which conditions and processing schemes the accuracy of water level measurements is higher with FFSAR rather than with UFSAR. We focus on Sentinel-3A, -3B and -6A missions during 2016–2023. Starting from the UFSAR products at a 20 Hz posting rate, we investigate the optimal choice for the Level 1 to Level 2 processing algorithms and parameters. We find that posting rates of 80 Hz for Sentinel-3 and 140 Hz for Sentinel6 are best suited for all inland targets. Our results indicate that the FFSAR, combined with the SAMOSA+ physical retracker, is likely to provide the highest accuracy and the largest number of observations for all targets, with averages of 15–20 cm in rivers, 6 cm in lakes, 40 cm in estuaries and 10 cm at the coast. The gain in accuracy of FFSAR over UFSAR is, for the same posting rate and retracker, smaller than 5 %, except for small targets with river widths smaller than 100 meters, areas smaller than 0.1 km2 and proximity to the coast. This study recommends FFSAR processing, although nadir-altimetry does not deliver longitudinal river profiles all along the river. Where water surfaces decorrelate quickly over time, the FFSAR accuracy may degrade, with a more pronounced impact over the open ocean compared to lakes.