Ultimately, product formulation requires being able to answer a simple question: what needs to be done in order to guarantee the desired product characteristics? The answer and how to get to it, however, is not as easy to state, and several approaches have been developed in the past to tackle it. This chapter presents key concepts regarding inverse design and provides an overview on some of these approaches, as well as how they relate to several of the previously addressed topics. As a manner of introduction to inverse design, the focus is placed on the partial least squares (PLS) model inversion. This is because PLS-based methods are relatively easy to apply, non-black-box techniques that provide dimensionality reduction through making use of existing collinearity in the data without significant loss of information, which is especially relevant for efficient model inversion when many highly correlated variables are involved