Uncertainty in soil parameters and variation in surcharge loading applied to the ground surface often make engineers overdesign for safety, which may lead to a cost inefficient design. In order to get a design, which is safe, insensitive to the uncertainties, and also cost efficient, the concept of robust geotechnical design (RGD) is adopted. In RGD methodology, robustness and cost are the main objectives to be optimized, while satisfying the safety requirement, by only adjusting the design parameters without reducing the existing uncertainties. In this paper, the aim is to develop an efficient and practical RGD for a design of a deep excavation supported by soldier pile anchor tieback system, using a single spreadsheet in Microsoft Excel. However, the analysis of complex supported excavation using numerical software such as Deepex 2015 within the RGD framework is computationally challenging. To make this RGD approach efficient and practical, the response surface procedure is adopted. In the design optimization, the signal-to-noise ratio (SNR) is taken as the robustness measure. The optimization between cost and robustness is performed in a single spreadsheet in Microsoft Excel using Solver. To locate the best compromise between these objectives, the minimum distance approach (MD) is adopted which can be implemented easily in a spreadsheet. Finally, the results are compared to the original design constructed in Taiwan.