Accessibility measurement methodology

The tools and calculations we use to measure accessibility.

Overview

The accessibility score expresses how well a website meets the WCAG 2.1 AA accessibility standards. The scale is 0–100 points, where 100 means no automatically detectable issues were found.

Automated testing is not enough

Automated testing catches roughly 25–30% of accessibility issues. A complete evaluation requires manual testing with real users.

How the score is calculated

The calculation starts from 100 points, from which points are deducted for the issues found. Each unique issue type is counted only once – if the same error is repeated in 20 places (e.g. missing alt text on images), it counts as a single issue, because it is one error in a template.

Weights by severity

Issues are divided into four severity levels. The more severe the issue, the bigger the deduction:

Critical issues:

  • 1st issue: 30 points
  • 2nd issue: 19 points
  • 3rd and further: 8 points each

Serious issues:

  • 1st issue: 15 points
  • 2nd issue: 10 points
  • 3rd and further: 5 points each

Moderate and minor issues: They do not affect the score, but their list is visible in every test.

Why decreasing weights?

For critical and serious issues we use decreasing weights. The first issue of a given severity has the biggest impact, further issues of the same severity have a progressively smaller effect. The reasons:

  • The first critical issue is a fundamental signal that something is very wrong
  • Further critical issues make the situation worse, but no longer as dramatically
  • Fixing the last issue brings the biggest reward – a motivation to finish the job

Calculation examples

  • No issues: 100 points
  • 1 critical issue: 100 - 30 = 70 points
  • 2 critical issues: 100 - 30 - 19 = 51 points
  • 3 critical issues: 100 - 30 - 19 - 8 = 43 points
  • 1 serious issue: 100 - 15 = 85 points
  • 1 critical + 2 serious: 100 - 30 - 15 - 10 = 45 points

Score interpretation

  • 90–100 points (Excellent): A minimum of automatically detectable issues. Good job!
  • 70–89 points (Good): Fix the critical and serious accessibility errors.
  • 0–69 points (Poor): The website has fundamental accessibility problems.

How fixes improve the score

Thanks to the decreasing weights: the closer you get to a clean website, the bigger the reward for each fix.

What we measure

We test against the WCAG 2.1 level AA standard using axe-core and HTML CodeSniffer tests in the PA11Y tool.

Examples of critical issues

  • Images without alternative text
  • Buttons without a label
  • Page zooming disabled (viewport)
  • Videos without captions

Examples of serious issues

  • Insufficient colour contrast
  • Links without distinguishable text
  • Form fields without labels
  • Incorrect heading structure

Examples of moderate issues

  • Missing page language
  • Incorrect ARIA attributes
  • Focus order problems

Examples of minor issues

  • Missing landmark regions
  • Recommendations for better accessibility

What we do not measure

Some aspects of accessibility cannot be tested automatically:

  • Comprehensibility of texts
  • Quality of alternative texts (only their presence)
  • Usability for users with cognitive impairments
  • Correctness of video captions

Excluded checks

Some checks are excluded from the evaluation because they relate to limitations of the testing tools, not to real accessibility barriers. For example, the "frame-tested" check reports that the tool could not test the content of an iframe (e.g. an embedded YouTube video). This is not an error of the website, but a technical limitation of the test.

These checks are shown in the detailed report for transparency, but they do not affect the resulting score.

Recommendations

  1. Start with critical issues – they have the biggest impact on users and on the score
  2. Fix systematically – the same error in many places = one fix in a template
  3. Test continuously – prevent issues from piling up
  4. Add manual testing – automated tests do not catch everything