How it works
Thorough by design, readable by anyone
What we test, how we test it, and how a list of thousands of warnings becomes a short list of things to fix.
1 · Scope
Every page, every size, every state
Problems hide in the parts of a page you only see after clicking. So we don't just load your pages; we use them.
Pages in scope
From your sitemap and menus, with templates grouped so one fix covers every page that shares it.
Desktop and phone
Each page at a desktop width and a phone width, because menus, layouts and touch targets change.
Interactive states
Menus opened, carousels advanced, tabs switched and dialogs shown. Carousels are paused so results are stable.
Forms, for real
Each form submitted empty, with mistakes, and correctly, on a test copy where possible, so error messages get tested too.
2 · Automated testing
Four engines, cross-checked
Engine versions are pinned and recorded in every report. When a version changes, the trend chart marks it.
axe-core
The most widely used open-source accessibility rules engine, from Deque. Strict about false positives.
IBM Equal Access
IBM's open-source checker with its own WCAG 2.2 rule set. Where it and axe flag the same element, confidence is high.
Audsey checks
Our own rules for problems the big engines skip but screen reader users notice: the same alt text on different photos, alt text that repeats the caption, hidden text that fails contrast, hidden fields with no name.
Accessibility tree
Reads what the browser hands to assistive technology: page titles, language, names of buttons and links, skip links, headings and landmarks.
3 · Testing by a person
What only a person can judge
Automated tools find roughly a third to a half of accessibility problems. The rest need someone to try the site.
- Keyboard only. Can you reach and use everything with Tab, Enter and the arrow keys? Is focus always visible? Does anything trap you?
- Screen reader. Every page with NVDA: does it make sense when heard rather than seen?
- Zoom and reflow. At 400% zoom and phone width, with increased text spacing, does anything overlap or disappear?
- Text over images. Contrast measured against the lightest and darkest pixels behind the text, in every state.
- Content. Do headings, link text, image descriptions and error messages say what they need to?
4 · Making sense of it
Thousands of warnings, a short list of causes
A raw scan of a small site can produce hundreds of warnings. Most are the same few problems repeated on every page.
Merge
Results from all four engines are matched to the same element, so one problem is counted once, with a confidence level.
Group
Instances are grouped by root cause: the same header button on every page is one issue, with every page listed.
Trace
Each group is traced to its source, whether that's the theme template, the page-builder widget or a plugin, with a link to edit it.
Decide
Alerts that need judgment are reviewed and marked fixed, verified as fine, or accepted with a reason. Nothing is silently ignored.
5 · The Accessibility index
One number, carefully defined
The Accessibility index runs from 0 to 100 and summarizes how well your site meets the WCAG criteria that apply to it.
- Each applicable WCAG criterion scores as met, partly met or not met, based on automated results and testing by a person.
- Level A criteria, the most basic barriers, weigh half again as much as Level AA criteria. Partly met counts for half.
- Criteria that don't apply to your site (no video, say) are left out rather than counted as passes.
- Criteria still waiting on a person's judgment are shown separately, not guessed.
The index is not a compliance percentage, and 100 is not a certificate. It means every criterion we tested was met on the pages and date in the report.
6 · Honest limits
What a report can and can't tell you
A report covers the pages listed, on the date tested, against the standard named. It is not a certification or a legal opinion, and it doesn't guarantee your site is free of every barrier or protected from claims. Testing by people who use assistive technology every day can find things any audit misses.
Third-party pieces, such as booking widgets, embedded maps, payment forms or reCAPTCHA, are tested as they appear on your site. When we can't fix them, your accessibility statement says so and offers another way to get the same thing done.
Read the full terms, or see how to read a report.
See it on your own site.
Tell us about your site and we'll show you what a baseline would cover.