To properly understand the social impact of direct-to-consumer genetic testing, we must consider the "sociotechnical architectures" of these technologies--how developers design and assemble the human and technical components of individual testing systems to perform specific functions. In particular, the way testing systems perform their main functions--providing access to testing, analyzing genetic material, and conveying test results--influence the technology's utility and the distribution of expertise in the medical system. I illustrate this concept by comparing two systems that offer single-nucleotide polymorphism analysis, a relatively new type of genetic testing. I conclude by exploring how policy officials and other decision makers might intervene in the development of sociotechnical architectures to maximize the benefits of genomic technologies.