This paper presents a methodology for integrating three tools for evaluating the performance of granular activated carbon (GAC) in the water treatment process in order to obtain the maximum benefit from each with the least cost and effort. They include mathematical models, bench-scale tests, and pilot column tests. This approach was applied in a 12-month study conducted for the City of Scottsdale, Arizona to develop GAC design criteria for a proposed 30-mgd surface water treatment plant. Methylisoborneol (MIB) and DBP precursor removal were the primary GAC treatment objective. Bench- and pilot-scale testing and computer modeling were used together as follows: rapid small-scale column tests (RSSCTs) were conducted to establish baseline pilot testing conditions and different GAC types (bituminous and lignite), and brands, were screened to select GACs for the pilot study, pilot test conditions for empty bed contact time (EBCT) and hydraulic loading rates (HLR) were selected based on RSSCT results; both bituminous and lignite GACs were tested (with and without ozone) in a 6-month pilot study to determine GAC performance for full-scale application with seasonal water quality changes and biodegradation; simulated distribution system (SDS) tests were conducted to establish seasonal TOC effluent targets based on TTHM and HAA5 formation; regression analyses were performed to fit the General Logistic Function model to the pilot-column data, and this mathematical model was used to evaluate a range of design parameters on either side of the test conditions to find the least cost design. Includes 7 references, figures.