We present results from a mixed methods study of screen reader use and switching behavior among people with vision impairments in India. We examine loyalty and experimentation with screen readers and find that the main drivers of adoption for early users differ significantly from the factors that drive continued use by advanced users. We discuss the factor that emerges as one of the strongest stated drivers of early adoption, TTS “voice” quality, particularly a “humansounding voice” as one of the key features differentiating free/open source (FOSS) products from more expensive proprietary products. While the initial preferences are driven by voice quality, application support becomes more important over time as users speed up their sound settings and become more comfortable with the resultant non-human-sounding speech. We discuss these findings from two theoretical perspectives – first, through the application of the economics of behavior switching, and second, vis-à-vis novice and expert approaches toward new product adoption. We argue that these findings further our understanding of initial user comfort related to assistive technology adoption, and the impact of early technology choices on long-term technology switching behavior.