Some aspects of design problems have been successfully delegated to machine learning models. Still, a more holistic view on data-driven design from a design perspective has not been developed. I want to propose a comprehensive model (a conceptual model, not a predictive machine learning model) that does not deny the complexity of graphic design processes and that helps to identify moments where data can be injected into the design process. Only then we can produce/gather “big” data, data that would be needed for computation.