A 3D-and-Excel-based predictive tool was developed to determine trunk spring movement for preventing recurrence of a noise problem. While effective, the tool could not completely explain measured results on the completed body unit (CBU). Since design data is used as the input, it was hypothesized that the difference between predicted and actual results was related to tolerance variation on the actual vehicle.Using Siemens® Variation Analysis software, the CBU was built and simulated virtually with tolerances using a Monte Carlo model. The study found that the hypothesis was correct; tolerance variation was fully responsible for the differences. In addition, the study also allowed accurate prediction of failure rates.