The aim of this study was to optimize the manufacturing process of an oil filter housing (made of PA 66 GF35) for automotive applications. Mass accumulations in the design of this component increased the cooling time in injection molding considerably. To improve the cooling, a copper-beryllium alloy mold insert with drilled cooling channels was installed.While copper-beryllium reduced the hot spot significantly – due to its thermal conductivity λ of 105 Wm-1K-1 - its wear resistance was too weak for the polymer used. Thus we searched for a solution where cooling performance is equal and wear resistance is superior. By using simulation, we analyzed two approaches to build up the insert: A) We filled up the core by copper and shielded it by a 3mm steel skin. B) We designed conformal cooling channels (Ø3.5 mm) in a Selective Laser Sinter (SLS)-able steel grade (λ = 23 Wm-1K-1). Modeling mold and process in simulation software Sigmasoft v. 5.1, we calculated first the heat transfer coefficients in the cooling channels using the CFD module. In the original design, the HTC was mostly far above 10000 Wm-2K-1, however, at bore crossings and in a dead end, the HTC fell locally to 700. Both the copper core insert and the conformal channel insert showed values > 12000. Second, we simulated injection molding cycles. Although in the insert socket region the cooling performance of copper core design and conformal channel design was significantly better than of the original design, the (cooling time determining) finger-like insert tips were less affected as neither copper nor conformal channel reach that region. Thus at the moment of ejection (after 75 s), where the recommended ejection temperature is <210°C, the relevant hot spot temperature at the part surface was 141°C, 213°C, 225°C for the original design, the copper core design, and the conformal channel design, respectively. In conclusion: If conformal cooling channels do not reach the region of interest, they are useless.