The objective of product design is to design a product whose performance is stable in the presence of noise factors which disturb the function of a product or process. We wish to find the optimum levels of controllable factors for the purpose of damping the effects of noise factors which degrade product performance stability. The traditional approach for achieving this goal has been based on a design - test - redesign philosophy. The purpose of this paper is to present a systematic method of optimizing the design to make it less sensitive to variations without adding cost based on the work of Dr. Taguchi. The paper dicusses the three stages of the design process: system design, parameter design and tolerance design. The philosophy of parameter design using orthogonal arrays and signal to noise ratio concept is developed. Steps for implementation of the method are given which can be used by product and process designers.