Defects in castings lead to non-conformities and affect productivity. Pull-down is a kind of defect occurring in castings. Several factors contribute to pull-down defects. Among those, pouring temperature, carbon equivalent and gating system are more significant. In the first stage, a set of process factors that were contributing to pull-down defect were identified. The identified factors were analyzed using 'Design of Experiments' approach. 'Signal-to-noise' ratio was estimated. Robust design factor values were estimated from the 'signal-to-noise' calculations. ANOVA analysis was done for robust design factor values. The interaction effects within factors were studied and the results were validated using 'Fisher's test'. In the second stage, optimized factor values were adopted in practical runs. It was identified that the optimized values had improved the acceptance percentage from 86.22% to 96.17%. The improved acceptance percentage had enhanced productivity of the foundry. © 2008 Elsevier B.V. All rights reserved.