In complex processes, the multiple properties that are measured from a raw material will have a combined impact onto the product's final quality attributes (FQA). The effects of the raw materials will be combined to those from the process and the scale of the operation. A new mathematical framework is developed to help understand and decouple the effects of raw materials, scale, and process conditions onto the FQA of a given product. Such framework allows the understanding of the process, the scale and the different control strategies available to operate the process since it decouples the effect of each of these onto the FQA. Ultimately it provides a mathematical framework to establish a multivariate specification purely driven by the FQA. The new technique is based on Latent variable models (LVM) coupled with optimization techniques. This proposed method can be used as the basis to quantitatively describe a design-space while taking into account the inherent variability in the FQA that is brought in by the process operation, the control system and the materials. The methodology is illustrated herein by applying it to establish a multivariate specification for the purchase of a polymeric excipient for the production of tablets.