Products with high product variety are often made in a manufacturing process (or a supply chain) consisting of multiple stages, with products taking certain features or “personalities” at each stage. The product may start as a common single engine. As the product moves along manufacturing process, more features are added, and the product assumes more identities of the final end product. When demands of the end products are variable from period to period, the production volumes of the intermediate stages in the manufacturing process are also variable. It is widely recognized that variabilities of production volumes may add cost to the process. This paper is motivated by our observations in industry, where some companies have reengineered the manufacturing process by reversing two consecutive stages of the process. Such changes could lead to variance reduction, thereby improving the performance of the process. We develop formalized models that characterize the impact of such changes: operations reversal. These models are used to derive insights on when such reversal would be advisable.