Estimating reading effort for screen reader users: a simple tool to support web accessibility awareness

Web navigation via screen readers is largely sequential and can impose a high cognitive and temporal effort on blind users. However, this aspect is often not very visible to developers, who rarely use screen readers themselves or have tools that make non-visual and keyboard-based reading experiences understandable. This work introduces a simple approach to estimating the reading effort associated with navigating a web page via a screen reader by applying structural metrics, such as word count, estimated reading time, heading structure, paragraphs, and interactive elements. In particular, we present a Python tool designed to automatically analyze web pages and return indicators intended to increase developers' awareness of how design choices impact the experience of screen reader users. The tool is not intended as a WCAG compliance checker, but rather as an accessibility awareness tool that translates technical aspects of a page into concrete estimates of the effort required during listening, taking into account different levels of screen reader user expertise. Two case studies on real-world websites show how apparently similar content and structures can result in significantly different reading loads. Finally, we discuss how the automatic estimation of reading effort and the structural analysis of content can also provide useful cues for conversational agents and large language models, supporting decisions about what to read aloud, how much to read, and in what order during voice-based interaction with blind users.