Back To App Documentation Accessibility Scanning How scanning works What a scan checks

What a scan checks

A scan loads each page you selected in a real web browser and runs automated accessibility tests against it. This article explains exactly what that covers, so you know how much weight to put on the result.

The standard

Scans test against the Web Content Accessibility Guidelines (WCAG) version 2.1, at levels A and AA. Level AA includes everything in level A, so together they are what people mean when they say "WCAG 2.1 AA". It is the standard most accessibility laws, policies and contracts point at.

WCAG 2.1 AA is made up of 49 individual requirements, called success criteria, that apply to web pages. Your results are organised around them.

How a page is tested

  • In a real browser. Pages are loaded the way a visitor's browser loads them, including content added by JavaScript. We wait for the page to finish loading before testing it.
  • At desktop width. Each page is tested at a desktop screen size. Problems that only appear on a phone sized layout, such as a menu that only exists on mobile, are not covered.
  • As an anonymous visitor. We see what a member of the public sees. Anything behind a login is not reached.
  • One page at a time, as loaded. We test the page as it first appears. We do not open menus, fill in forms or click through multi step processes.

What automated testing finds well

Missing text alternatives on images, form fields with no label, text with too little contrast against its background, missing page language, empty links and buttons, broken ARIA attributes, and a long list of similar problems that can be judged from the page's code. When an automated test reports one of these, it is a real, objective failure.

What it cannot find

Around half of WCAG needs human judgement. An automated test cannot tell whether alternative text actually describes the image, whether a page can be used with a keyboard alone from start to finish, whether captions match what is said in a video, or whether error messages make sense. So a scan with no findings is good news, but it is not the same as a site that meets the standard.

That gap is what an Expert Manual Review is for. See Automated results and expert review.

What is not tested

  • PDF and other documents. Links to PDFs are listed for you, but the documents themselves are not tested. See PDF documents on your site.
  • Content on other websites, including embedded content served from somewhere else.

Next

Read Automated results and expert review. It is the most important article in this space.

Was this page helpful?