The formation of different mesoscale networks in multiwalled carbon nanotube (CNT) systems that are realized by the mixing of CNTs of varying lengths and length dispersities is reported. By this mixing process, we introduce competing length scales; hence, we alter the self-organized packing that contributes to the synergistic effects on the functional properties of the networks. The experimental findings show a gradual change of volume fraction and aspect ratio as well as a 2-fold increase in electrical conductivity for such networks at certain specific compositions, hence an average aspect ratio. Quantitative large-area scanning transmission electron microscopy (STEM) imaging indeed revealed the existence of such mesoscale packing distribution variations. These packing observations suggest that these optimized networks of CNTs fit into an electrical conduction model that attributes its behavior to the formation of conduction knots, i.e., high-volume-fraction regions of relatively short CNTs that are connected by relatively long CNTs. If these conduction knots occupied by many short CNTs are distributed evenly with their nearest-neighbor distance being close to the average length of the CNT population, then the total electrical contact resistance in the conduction path will be effectively minimized. This study shows that optimized macroscale functional material properties can be designed into the initial colloidal dispersions by understanding and thus tuning the self-organization behavior of colloidal matter.