Design of Experiments (DOE) techniques were applied to improve the thermal performance of a gas oven, aided with a Computational Fluid Dynamics (CFD) study in concept selection phase of Product Design Process PDP. Typical approaches to these types of activities involve developing all possible combinations of geometries changing one variable at a time, analyzing them with CFD, and predicting the main effects over the parameters surface heat flux and pressure drop, which in this application are the system responses. The develop plan for the team was to generate an script that allowed analyzing the sample space of the process variables and analyze much geometry configurations to study the geometric parameters. By utilizing DOE techniques the number of geometries was strategically reduced to 29. The metamodel of the system was obtained as a function of the input variables, with a statistical adjustment of 86.78% for the surface heat flux and 78% for the pressure drop. Subsequently we found the geometry that improves the performance of the product using the multiobjetive optimization method "Normal Boundary Intersection -NBI".