WCAG
Accessibility Testing: How to Test Your Website Properly
Contents
Before you can fix accessibility barriers, you need to find them. Testing is the first and most important step towards an accessible website—and also the step where most mistakes are made. “Testing a website for accessibility” is often reduced to a quick automated scan, which is misleading because a tool can only detect part of the picture.
This article explains what automated testing can and cannot do, how manual testing and assistive technologies cover the remaining requirements, and how test results lead to genuine remediation—instead of a green checkmark that proves nothing.
Why testing is the first step
You cannot fix problems you do not know about. The WebAIM Million report demonstrates how widespread these issues are: in 2024, automatically detectable WCAG errors were found on more than 96% of the home pages tested. The problem is therefore almost universal, and the first step towards solving it is an honest assessment of the current situation.
The right approach is important: testing is not a one-time activity but part of a continuous cycle of testing, fixing, and retesting. How to Make a Website Accessible explains how the complete process fits together.
Automated testing: fast but limited
Automated tools are a good starting point. They reliably identify errors that follow clear, programmable rules across many pages, such as missing alternative text, insufficient contrast, missing form labels, and structural problems.
However, their limitations are fundamental—and honesty is more useful here than marketing figures. Measured against the WCAG success criteria, automated tools can meaningfully assess only about one-third of them—approximately 16 of the 50 Level AA criteria. The rest require human judgement.
An important distinction must be made: when measuring the total number of individual issues detected instead, the percentage can be higher. Deque states that its axe-core tool can identify approximately 57% of accessibility issues because common problems, such as contrast failures, often occur many times and can be detected automatically. However, this does not mean “57% conformant.” It only means that the tool detects many individual issues while missing entire categories of requirements.
The reason is simple: a tool cannot determine whether alternative text is meaningful, whether an image is decorative or informative, whether the focus order is logical, or whether text is clearly written. For a detailed explanation of what tools can assess and where their capabilities end, see Automated WCAG Testing.
Another warning sign: incorrectly implemented ARIA does not improve a website—it can make it worse. On average, pages using ARIA contain more detected accessibility errors than pages without it. This is precisely the type of problem created by overlays and rushed quick fixes. Learn more in Accessibility Overlays.
Manual testing: what tools cannot see
Manual testing allows a person to assess the requirements that demand judgement—which represents the larger proportion of the criteria:
- Are alternative texts meaningful, rather than merely present?
- Is the focus order logical and understandable?
- Are the content, labels, and instructions clear?
- Do interactive processes—such as forms, filters, and menus—actually work for everyone?
Manual testing is not a “nice-to-have.” Without it, no reliable statement about conformance is possible.
Testing with assistive technologies
The most realistic impression comes from using the website in the same way as the people affected:
- Keyboard testing: Operate the entire website using only Tab, Enter, and the arrow keys. Can you reach everything? Is the focus always visible? Are there any places where you become trapped?
- Screen-reader testing: Navigate through the website with a screen reader. NVDA is free on Windows, VoiceOver is built into macOS and iOS, and TalkBack is included with Android. Is everything announced meaningfully? Are the headings and landmarks correct?
- Real users: Whenever possible, testing with people with disabilities provides the most meaningful assessment.
From test results to remediation
Testing is not an end in itself. The result is a list of genuine problems—and those problems must be fixed in the source code, not concealed. The process is straightforward: test, fix the code, retest, and gradually improve the measurable results.
This is where an honest scanner helps: it identifies genuine source-code problems and makes progress measurable through repeated scans—instead of simulating a green checkmark that would not withstand proper assessment. Welcoma’s upcoming WCAG scanner is designed for exactly this purpose: revealing what is really there so that you can fix it.
Overview of the tools
- For a quick initial assessment: axe by Deque, WAVE by WebAIM, Lighthouse built into Chrome, and Pa11y for automated testing pipelines. They provide an initial list of errors within minutes.
- For manual testing with assistive technology: NVDA, VoiceOver, and TalkBack.
The key point is that none of these tools replaces the others. Automated tools provide a quick overview, manual testing provides depth, and assistive technologies reveal the real user experience. A single “green” result from one tool is not proof of conformance—anyone claiming otherwise is selling appearance instead of substance.
Frequently Asked Questions
Is a quick online test sufficient to assess my website?
No. An automated scan covers only about one-third of the WCAG criteria. A reliable assessment also requires manual testing and tests with assistive technologies.
Which free tools can I use?
WAVE, Lighthouse, and axe are free for initial automated checks. The NVDA screen reader for Windows is also free, while VoiceOver and TalkBack are built into their respective operating systems.
How often should I test?
Regularly and after every major change. Accessibility is an ongoing process—new content and features can introduce new barriers at any time.
Updated: . This article gives a general overview. Not legal advice.