The valorisation of waste or by-products in Portland clinker production is a promising alternative for developing sustainable cements. The complexity of the chemical reactions during clinkering demands an adequate dosing method that considers the effect of feedstock impurities to maximise the potential substitution of natural resources by waste or by-products, while guaranteeing the clinker reactivity requirements. This study proposes a raw meal proportioning methodology for optimising co-processing of natural feedstocks with alternative raw materials in clinker production, intending to reduce the content of natural raw materials needed, while promoting an optimal clinker reactivity. A thermodynamic modelling sequence was developed considering the variability of raw materials composition and heating temperatures. The model was then validated by comparing simulation outcomes with results reported in previous studies. An experimental case study was conducted for validation of the proposed method using a spent fluid catalytic cracking catalyst (SFCC), a by-product from the oil industry as an alternative alumina source during clinkering. The modelling simulations indicated that substitution of natural feedstocks by 15 wt% SFCC promotes the formation of reactive clinkers with more than 54% tricalcium silicate (C3S). Mixes with the potential to form the highest C3S were then produced, and heating microscopy fusibility testing was applied for evaluating the clinkers’ stability. The main factors governing the reactivity and stability of the clinker phases were the melt phase content, alumina modulus, and formation of C3S and dicalcium silicate (C2S). The self-pulverisation of clinker during cooling was observed in selected mixes, and it is potentially associated with high viscosity and low Fe content in the melt phase. The proposed framework enables optimisation of the dosing of raw meals containing alternative alumina-rich feedstocks for clinker production and allows a deeper interpretation of limited sets of empirical data.