Radiation efficiency is a key performance indicator for multi-standards frequency agile electrically small antennas (ESA) that are mounted on wireless IoT sensors. One of the techniques to estimate it, consists to integrate, over all the angular directions, the gain measured in the far field condition. The gain-comparison method is usually implemented in the CEA LETI test bench, which requires an accurate knowledge of the standard horn gain. The introduction of a new RF-optical link to remove coaxial cable perturbation on ESA radiation, in our test bench has raised the opportunity to proceed to an error budget analysis. This paper delivers the main results of this study where the impact of several parameters such as the optical fiber movement, the horn positioning, the received power level, and the chamber imperfection have been evaluated. The three antennas method (one Vivaldi and two TEM standard horns) is carried on to estimate the complex transfer function of the three antennas. The overall goal is to estimate the detailed uncertainty analysis of the ESA efficiency measurement over a large frequency band. This work aims to identify the most impacting effects on uncertainty and to initiate the discussion with the AMTA community how to decrease them. © 2016 AMTA.