A robust model-based iterative feedback optimization methodology for the steady-state optimization of chemical process operations was introduced in a previous work (Cheng, J.-H.; Zafiriou, E. Ind. Eng. Chem. Res. 2000, 39, 4215) as a complementary approach to conventional realtime optimization (RTO) methods to improve plant operations without requiring cumbersome model updating. In this paper, the results analysis component that is an integral part of an RTO system is developed for use with this methodology to evaluate the inherent variability of the optimization results transmitted from the plant measurements. In this way, the method is able to distinguish the effect of model-plant mismatch from that of measurement noise, and only optimization results that represent meaningful changes are implemented as the new set points, thus reducing unnecessary and profitless corrective actions. Furthermore, the iterative nature of the methodology allows for the development of a condition for guaranteed improvement over iterations, which can be used to quantify, a priori, the limits that measurement noise imposes on achieving optimal operation. The effectiveness of the method is demonstrated on a simulated CSTR process.