Today, CFD is commonly used by engineers for the rheological design of extrusion dies. However, the design process per se still depends on the craftsmanship of the designer. To overcome this dependency of manual labour, automatic optimisation methods were introduced into the field of die design. These numerical optimisers use a lot of simulation data to determine the “optimal” design candidate. Because of the memory demanding nature of CFD data, usually only the end results of such optimisations are stored permanently, all intermediate runs are discarded. However, each dataset holds a valid relationship between quality criteria such as the outlet velocity distribution and the die geometry, melt properties, and operating point. This information could become handy, if you must design a geometrical similarly die, or you want to optimise the die for a differing operating point. We present an approach how intermediate optimisation runs can be used to generate metadata characterising a design candidate (incorporating die geometry and melt rheology) in a way, that it can be used for preliminary design studies for geometrical related dies. This is achieved by extracting aggregated data from the CFD post-processor and storing it in an independent database. The presented showcase demonstrates that the method is even possible with commercial CFD codes using Ansys Fluent as an example.