Latest / Software Testing with Fexingo: QA, Automation, and Reliable Software Engineering / Why Accessibility Testing Is Not Optional
Transcript
- Lucas: So, Luna — when was the last time you tested a screen reader against a feature you shipped? Luna: Honestly? Not recently. We rely on our accessibility team for that. But I think a lot of teams do — they treat it as a specialist concern. Lucas: Right, and that's exactly the problem. Accessibility testing is still seen as this optional, separate activity — something you do after the fact, almost like a compliance checkbox. But the data says otherwise. A 2023 WebAIM study found that 96.8 percent of the top one million homepages have detectable WCAG failures. Luna: That number is wild. Almost every major site is failing in some way. What's the most common violation? Lucas: Low contrast text — roughly 86 percent of homepages have that. Then missing alternative text for images, missing form labels. These are the basics, the stuff that's been in the guidelines for years. And yet they're everywhere. Luna: So why aren't teams catching this? I mean, there are automated tools. You can run axe-core in your CI pipeline in minutes. Lucas: Exactly. And that's the part I find frustrating. The tools exist, they're free, and they integrate into any CI/CD workflow. But teams still treat accessibility as a manual QA step that happens right before release — if at all. The result is that you fix maybe the top ten issues, but the deeper problems — like screen reader navigation flow or keyboard traps — never get caught. Luna: I've definitely shipped a feature where a user couldn't tab through a modal. We had to hotfix it two days later. That's embarrassing. Lucas: Right, and that's exactly the kind of bug that automated unit tests don't catch. You need either a manual screen reader test or a dedicated tool like ja11y or AccessLint in your CI. The key is shifting left — test for accessibility as early as you test for functionality. Luna: But there's a pushback I hear a lot: 'We don't have the budget for that.' Or 'Our users don't have disabilities.' What do you say to that? Lucas: I say look at the numbers. According to the CDC, one in four US adults has some type of disability. That's about 61 million people. And that's not even counting temporary impairments — someone with a broken arm, a parent holding a baby, a user in bright sunlight. Accessibility isn't a niche concern; it's a universal design principle. Luna: And there's the legal risk too. ada related web accessibility lawsuits have been increasing every year. In 2024, there were over 4,600 filed in federal courts. That's not a small threat. Lucas: Exactly. But I don't want to frame this just as risk avoidance. There's a real business upside. For instance, a fintech startup I know embedded axe-core into their pull request workflow. Every PR that introduces a new component gets flagged if it has contrast issues or missing ARIA labels. Within three months, their accessibility defect rate dropped by 70 percent. And because their site was more usable, they saw a measurable improvement in conversion rates — especially on mobile. Luna: So they basically treated accessibility as a feature, not a bug. That's smart. Did they also do manual testing, or was the automation enough? Lucas: Automation catches about 30 to 50 percent of WCAG violations, depending on the standard you're measuring against. So it's a great first pass, but you absolutely need manual testing with real screen readers — like NVDA on Windows or VoiceOver on Mac and iOS. The combination is powerful. Automation gives you constant scanning; manual testing gives you the human judgment to assess whether a screen reader user can actually complete a task. Luna: I've heard that one common mistake is focusing only on the code and not the content. Like, you can have perfect ARIA labels, but if the copy is confusing, the user is still lost. Lucas: That's a great point. Accessible design isn't just technical — it's also about clear language, logical heading structures, and meaningful link text. A button that says 'Click Here' is bad for everyone, but it's especially bad for someone using a screen reader who's tabbing through links. They need context. Luna: So if a team wanted to start today, what's the single most impactful thing they could do? Lucas: Run an automated scan on your top twenty user flows. Use a tool like axe DevTools — it's a browser extension, free, and it gives you a clear report. Then fix the critical issues — missing alt text, low contrast, missing form labels. That alone will get you past the low-hanging fruit and probably eliminate the most common user complaints. Then set up a CI check so those issues never regress. Luna: That's not a huge lift. A few hours of work, and you've already made your product more inclusive. So why do so many teams still avoid it? Lucas: I think it's a combination of not knowing where to start and the misconception that it's expensive. But honestly, the cost of not doing it is higher — in legal fees, lost customers, and brand damage. And if today's conversation gave you something usable to bring back to your team, that's exactly the kind of thing that makes all this worth doing. Luna: Definitely. And the way we keep making these episodes ad-free is listener support. If you found this useful, you can buy me a coffee at buy me a coffee dot com slash fexingo. It's a simple way to help us keep going. Lucas: Right, that's the only reason we can do deep dives like this without sponsors. So if you're able, that link is buy me a coffee dot com slash fexingo. Now, back to the tools — Luna, one thing I didn't mention is that some teams pair accessibility checks with visual regression testing. Because accessibility issues often overlap with layout changes. Luna: Oh, that's clever. So you could catch a contrast violation as part of your visual diff. That's a nice way to embed it into an existing workflow. Lucas: Exactly. And that's the mindset shift I think we need: stop thinking of accessibility as a separate workstream and start thinking of it as a core quality attribute — like performance or security. You wouldn't ship a feature that crashes in Chrome, so why ship one that's unusable with a screen reader? Luna: Because it's invisible to most of the team, right? Unless you're actively testing with assistive tech, you don't feel the pain. That's the real challenge — building empathy into the process. Lucas: Yeah, and the best way to build that empathy is to have developers spend thirty minutes with a screen reader. Just try to navigate your own site. It's eye-opening. And once you've done that, the automated checks suddenly feel like a no-brainer. Luna: Alright, I'm convinced. I'm going to add an accessibility check to our next sprint. And maybe even schedule a screen-reader workshop for the team. Lucas: That's the spirit. And if you do, let me know how it goes — I'd love to hear what you find.