The spatial resolution of energy system optimisation models (ESOMs) is often compromised for computational performance. Reducing the spatial resolution impacts the least-cost solutions when optimising the generation capacity and transmission capacity of the ESOMs with high penetration of variable renewable energy sources. Previous studies show that the two main effects of spatial resolution reduction are the increase in solar and wind expansion capacity and the decrease in transmission capacity expansion. This paper introduces a targeted method of defining regions during spatial scale reduction by using the max-p-regions clustering algorithm to aggregate similar areas into regions. The attributes used to determine the similarity between areas in the max-p-regions method are population; wind and solar resource potential; and pumped-hydro storage capacity. Two alternative spatial resolution reduction methods were used to compare how the effects of spatial resolution impacted the optimisation results of the ESOMs. Evaluation results for three country groups showed that using the max-p-regions method to define regions caused the two effects of spatial resolution reduction to be generally lower than using national jurisdictions. For the case studies Germany and Spain, the results showed that the max-p-regions method identified sets of regions, which were less impacted by the two effects of spatial resolution reduction.