The computational feasibility of many systems with large degrees of freedom such as chemically reacting systems hinges on the reduction of the set to a manageable size with a minimal loss of relevant information. Several sophisticated reduction techniques based on different rationales have been proposed; however, there is no consensus on the best approach or method. While the search for simple but accurate schemes continues, the classical quasi-steady state assumption (QSSA), despite serious shortcomings, remains popular due to its conceptual and computational simplicity. Invoking the similarity between a reduced invariant manifold and a streamline in fluid flow, we develop an advanced QSSA procedure which yields the accuracy of more complex reduction schemes. This flow-physics inspired approach also serves to reconcile the classical QSSA approach with recent methods such as functional equation truncation (FET) and intrinsic low dimensional manifold (ILDM) approaches.