Look, accessibility* is great and important to many people, and if you can afford to prioritize it early on in your product development, that's great - but it is not productive to treat commercial products like critical infrastructure, ignore the costs of implementing accessibility features (common tooling and standardization make it much easier, thankfully), and exaggerate the benefits.

The article starts off with the over-cited and misunderstood 1/6 figure, which comes from the WHO's claim that 1/6 of the world suffers from a "significant disability". I couldn't find a breakdown of what the WHO considers a significant disability in the first place, but the CDC in the US lists the following six disability categories in one of its studies: hearing, vision, cognitive, mobility, self-care, and independent living. How many of these are actually relevant in software? I have never once heard calls to make software more accessible to people with cognitive disabilities, and mobility/self-care/independent living categories are simply irrelevant to the overwhelming majority of software use cases. So no, you're not growing your audience by 1/6 by implementing screen-reader support for your website. If your product is growing, you're better off focusing on internationalization (which, technically, is a form of accessibility, but obviously not the subject of the article).

* Do we really need to abbreviate "accessibility" everywhere? It's just nine more characters than "a11y" and - speaking of accessibility - is more readable!