Accessibility Testing

For a free quotation please email us, telephone 0870 855 3370 or fill in the form on our Contact page

Technical Testing

The starting point for most projects is usually the W3C's Web Content Accessibility Guidelines (WCAG), which are a set of technical criteria to assess how accessible a website should be. We have developed a range of manual testing techniques for each of the criteria, and use lightweight tools where they improve efficiency.

User Testing

Technical testing only indicates what the level of accessibility should be, not what it actually is. To determine the actual accessibility of a website it is necessary to test it with a range of users with different disabilities and assistive technologies such as screen readers and screen magnifiers.

A major limitation of manual and automated tests is that they test each page in isolation, but typically this is not how people use a website. Most website use is based around tasks, so we conduct scenario based tests in facilitated sessions.

Expert Review

When the time or budget available is limited, valuable results can be obtained by means of expert review in lieu of user testing. This is conducted by a tester familiar with assistive technologies and the issues encountered by disabled users of these technologies.

Automated Tools v Manual Testing

Several tools are available to automate accessibility testing but we concur with the W3C whose website states "No single evaluation tool yet provides comprehensive information or captures all problems with regard to the accessibility of a site".

In fact tools can only make an assessment of about 30% of the WCAG criteria so a significant amount of manual work is required if testing is to be anything other than superficial.

Manual testing is a highly technical task involving both code inspection and testing via the user interface. Testing with a variety of user agents (graphical browsers, text browsers, screen readers etc) is recommended because some of these can exhibit adverse behaviour that cannot be predicted from testing with a single browser.