Many stakeholders today consider power tower systems to be the most promising CSP technology for the future. Therefore much effort is spent to improve individual components and subsystems. While these are valuable steps towards cost reduction, it is also of great importance to perform an overall, i.e. ‘holistic’, system layout and optimization process. Still, even today heliostats are often characterized and even compared using the single parameter ‘cost per square meter’, which is not sufficient. More factors like optical and tracking accuracy, shape, structural deformation under operation loads and maybe even power consumption have to be factored in to allow for a meaningful comparison.In the paper recent activities targeting at a more holistic power tower system optimization are described. For the overall layout process of heliostat field, tower and receiver the complete system is modeled using SAM or an in-house tool based on SolTrace and Matlab to determine investment cost, annual electricity generation and resulting levelised electricity costs. By doing so, different heliostat types and field layouts can be directly compared using LCoE as a reasonable figure of merit.This logical approach to identify the optimum heliostat design is described and illustrated using two generic heliostat designs and the respective field layouts as an example; a recently developed updated tower cost curve is used. It is found that using specific heliostat costs (i.e. $/m2) alone as a figure of merit to assess different heliostat designs can be misleading, because heliostats characterized by lower specific investment costs, but also lower optical and tracking accuracy, can lead to higher levelised electricity costs, which means the real objective of power plant optimization is not reached by looking at heliostat costs per square meter alone.