Polymer nanocomposites (PNCs) hold great promise as future lightweight functional materials processable by additive manufacturing technologies. However, their rapid deployment is hindered by their performance depending strongly on the nanoparticle (NP) spatial organization. Therefore, the ability to control nanoparticle dispersion in the process of PNC preparation is a crucial prerequisite for utilizing their potential in functional composites. We report on the bulk processing technique of tailored NP spatial organization in a model glass forming polymer matrix controlled by structural and kinetic variables of the preparation protocol. Namely, we studied the impact of solvent on the NP arrangement, which was already known as a tuning parameter of the solid-state structure. We emphasized the qualitative differences between "poorly dispersed" NP arrays, which, by combination of rheological assessment and structural analysis (TEM, USAXS), we identified as chain bound clusters and aggregates of either thermodynamic or kinetical origin. They are characterized by substantially distinct formation kinetics and mismatched properties compared to each other and individually dispersed NPs. We quantitatively linked all the currently observed types of NP dispersion with their rheological properties during the solution blending step and the amount of polymer adsorption and depletion attraction. We propose the ratio of NP-polymer and NP-solvent enthalpy of adsorption as a parameter capable of the quantitative prediction of NP arrangement in systems similar to our current model PNC. Finally, we bring forth the comparison of glass transition temperatures to further demonstrate the importance of NP spatial organization in PNCs.