Increasing interest in hydrogen as a clean and efficient energy carrier has driven the need to improve its production and distribution methods. Of particular interest is the natural gas reforming process which currently accounts for 48% of hydrogen production in the United States. Existing natural gas delivery infrastructure and challenges associated with hydrogen delivery suggest distributed (~1,500 kg/day) natural gas reforming may be the most effective production method for the transition into a hydrogen economy. A key challenge hindering the use of distributed steam reforming is the high hydrogen cost compared to central production plants (~750,000 kg/day), which benefit from economy of scale. Improved hydrogen purification methods at the back end of the reforming process that scale to expected distributed production levels will help curtail this disparity. Highly selective palladium (Pd) alloy membranes have shown promise as an efficient separation method by producing nearly pure hydrogen at high recoveries. While the mechanisms of mass transfer through the composite Pd membranes are well understood, mass transfer resistance by the porous support has been largely overlooked, as much of current research is directed toward the Pd itself. Depending on support pore sizes, the supports can dominate the mass transfer resistance, highly influencing the overall performance of the membrane. Additionally, characterization of the porous supports is required as variations in the properties of commercial porous stainless steel (PSS) supports may also greatly influence not only mass transfer resistance by the support, but also required Pd thickness. When used in the full steam reforming process, the influence these factors have on membrane performance may alter the efficiency of hydrogen production. This work addresses these topics through development of a model to predict membrane performance as a function of membrane length and support pore size. Results from this model show that the support can greatly control mass transfer through the membrane depending on the pore size configuration of the support. By using a stepwise graded support that minimizes mass transfer resistance while maintaining a fine surface layer for Pd deposition, membrane performance is improved. The impact of support mass transfer resistance is then translated to a full steam reforming process model to show the overall reactant and utility use as a function of membrane recovery. Although the energy intensity of steam reforming largely overshadows reductions in reactant and utility usage from improved membrane performance, the opportunity to later realize these benefits is presented. Lastly, the bubble-point test is used to assess variations in porous stainless steel (PSS) filtration grade of several as-ordered PSS discs. Results show that while there is slight variation in filtration grade between each disc, they are all double the ordered filtration grade. While a larger than expected filtration grade does not directly worsen mass transfer, it may imply that surface pore sizes would be larger than expected, directly impacting required Pd thickness. A modification to the bubble-point test is discussed that may characterize surface pores during future research.