Latest / Software Testing with Fexingo: QA, Automation, and Reliable Software Engineering / How Accessibility Testing Uncovers Invisible Usability Bugs
Transcript
- Lucas: So earlier this year, a major e-commerce platform — I won't name them, but you've definitely bought something from them — rolled out a redesign of their checkout flow. Luna: I think I know the one. It was all over Twitter for about a week. Lucas: Exactly. The redesign was sleek, mobile-first, all that. But within days, the accessibility team discovered that nearly 40 percent of the checkout process was completely invisible to screen readers. Luna: Forty percent? That's not a small edge case. That's half the purchase flow. Lucas: Right. And the thing is, their unit tests passed. Their integration tests passed. They had great code coverage. But none of those tests checked what happens when a blind user tries to navigate that form using a screen reader. Luna: So they had a classic gap: coverage metrics that looked great, but real-world usability that was broken for millions of users. Lucas: Exactly. And that's what today's episode is about — why accessibility testing isn't just a nice to have compliance checkbox. It's a testing discipline that uncovers bugs that affect everyone, not just people with disabilities. Luna: I love that framing. So where do we start? Automated checks? Lucas: Let's start with what automated tools actually catch. Tools like axe-core, Lighthouse, or WAVE — they're great for finding things like missing alt text, insufficient color contrast, or missing form labels. But studies show they only catch about 30 percent of all accessibility violations. Luna: Thirty percent? That's lower than I thought. So the remaining 70 percent require human judgment? Lucas: Right. Automated tools can't tell if a screen reader user can actually complete a task. They can tell you an element has an aria-label, but they can't tell you the label makes sense in context. That requires manual testing with real assistive technologies. Luna: It starts with understanding the Web Content Accessibility Guidelines, version 2.2. That's the current standard. WCAG 2.2 has four principles: perceivable, operable, understandable, and robust. Each principle has specific success criteria at levels A, AA, and AAA. Lucas: Yeah, that's the baseline. But testing those criteria isn't just following a checklist. For example, testing keyboard navigation — you have to tab through every interactive element and make sure you can complete the entire workflow without touching a mouse. Lucas: I've seen teams automate some of that with tools like axe-core in their CI/CD pipeline. They run a scan on every pull request. But the manual testing — that's often done once per sprint, or worse, once per release. Luna: Exactly. And there's a practical way to do that: add accessibility checks into your component testing. When you write a unit test for a button, also check that it has an accessible name. When you test a form, verify that error messages are announced by screen readers. Luna: Testing Library has built-in accessibility queries like getByRole and findByLabelText. They encourage you to find elements the way a user does — by their accessible name, not by CSS class. And Jest-axe lets you assert that a component has no accessibility violations. Luna: Right. And then for end to end tests, you can use Playwright or Cypress with axe-core integration to scan every page after critical user actions. But even then, you need real screen reader testing. Luna: Some teams schedule a monthly pair-testing session where one engineer uses a screen reader — NVDA or VoiceOver — and another documents issues. They go through the three or four most critical user journeys. It's surprisingly fast once you get used to it. Luna: Absolutely. But that's exactly the point — it reveals assumptions your sighted team made. Like that checkout example I mentioned: the developers had used custom dropdowns that weren't keyboard accessible. A screen reader would just skip right over them. Luna: Exactly. Accessibility fixes improve usability for all users. Curb cuts, captions, high contrast — they were designed for specific disabilities but ended up helping everyone. Lucas: I'd say: run an automated scan on your top five user flows today. Then, within the next sprint, do one manual round of keyboard-only testing on those same flows. You will find bugs. I can almost guarantee it. Luna: That's where risk-based prioritization comes in. A bug that prevents a user from completing a purchase — that's a showstopper, regardless of whether it's an accessibility issue. It's a revenue blocker. Luna: All good reasons. But honestly, the strongest reason for me is that it makes you a better tester. Accessibility testing forces you to think about how your software actually behaves, not just how it looks in a browser. Luna: That's a great note to end on. If today's conversation gave you something usable, consider supporting the show. It's the reason we can keep these episodes ad-free and focused on practical testing advice. Lucas: Yeah, if it was worth a coffee to you, the link is buy me a coffee dot com slash fexingo. No pressure, seriously. Every bit helps us keep the lights on. Luna: And now back to the topic: one more practical tip. If you want to get started with screen reader testing today, download NVDA if you're on Windows, or use VoiceOver on Mac — it's built in. Pick one critical flow and try to complete it with your eyes closed. Lucas: That's a powerful exercise. It only takes a few minutes, and it will show you things your test suite never caught. Luna: Alright, that's all for today. Thanks for listening.