Initial product designs contain some parameters mapped by study and it is extremely important a full model understanding of process. Design of experiments methodology (DOE) shows some approaches for this goal and one of them is factorial design. However, a full factorial design is not recommended for more than four parameters, which would demand lots of testing run and initial budget for development projects is limited. This problem gets worse when it is desired to optimize two contradiction parameters, e. g., light weight and high strength material. This work deals with comparison of some initial design of experiments aiming to improve basic new product or process knowledge by as few as possible runs. It was performed some variations of fractional factorial design and a Plackett-Burman method and compared among them and their impacts in analysis. The analysis quality is very important for selecting parameter for optimization and a poor experimental result could impact in final product quality.