A framework was developed for preliminary evaluation of the relative media replacement costs of three alternative column configurations used for adsorption systems with two vessels, such as those serving small systems. The media replacement cost is the cost of fresh media and the replacement service cost (including transportation, labor, and other non-material costs). Cost normalization methods were developed in part based on the data from US EPA Arsenic Treatment Technology Demonstration Program. Adsorption equilibrium and kinetics were modeled using the PSDM model and breakthrough curves were normalized using the target effluent to influent concentration ratio (C/Co) and the mass transfer zone fraction (%MTZBT). Two factors were found to be important for the relative replacement cost of each configuration - the frequency which at least one column needed replacement of media, and the cycle replacement cost (CRCost) which is a combination of the fresh media cost and the replacement service cost. The lead-lag configuration has the lowest annual replacement cost at low target C/Co, high %MTZBT, and high CRCost ratios. The parallel configuration performs better at high target C/Co, high %MTZBT, and high CRCost ratios. Although the single configuration (two columns operated in tandem and replaced simultaneously) has higher media consumption compared to lead-lag and parallel, it can result in the lowest replacement cost at short %MTZBT and very low CRCost ratios due to savings in the replacement service cost.