In the track to fuel efficiency increase and green house gases emissions reduction, systemic friction reduction in mechanical components starts to play a big role in the complete automotive system design. When thinking on design for friction efficiency, most of the common applied mechanical components still have enough room for improvement, so as to greatly impact the whole mechanical efficiency. Following the results obtained by a Design and Analysis of Computer Experiments - DACE investigation on the most influencing friction torque parameter of roller bearings, different designs of raceway profiles were examined in order to reduce the bearing friction, i.e., increase its efficiency. In this work a DACE methodology was applied alongside a friction model for the roller bearings to evaluate the effects of the variables and its interactions. For simulation studies a surrogate model is constructed and based on the results an optimal level of factors is found for the friction torque reduction.