What Does an Accessibility Score Actually Mean?

Every accessibility scanner gives you a score. Lumi gives you a score out of 100. Lighthouse gives you a percentage. WAVE shows error counts.

But what do these numbers actually mean? And more importantly, what should you do with them?

How Lumi calculates your score

Lumi uses a penalty-based scoring system. You start at 100 and lose points for each accessibility issue found, weighted by severity:

• Critical issues deduct the most. These are barriers that completely prevent users from accessing content or functionality.
• Serious issues deduct significantly. These make it very difficult for users with disabilities to use your site.
• Moderate issues deduct less. These create friction but don't block access entirely.
• Minor issues deduct the least. These are technical violations that have limited practical impact.

The score is capped - it can't go below 0 and a single critical issue can significantly impact your score. This is intentional: a single keyboard trap matters more than twenty minor contrast issues.

What different scores mean in practice

90-100: Strong foundation. Your site has few or no critical and serious issues. Automated scanning is finding minimal barriers. You likely have a team that thinks about accessibility during development. Next step: manual audit to catch what automation misses.

70-89: Common range for sites that have done some accessibility work. Expect a handful of serious issues and some moderate ones. These sites typically have good intentions but gaps in implementation. Most issues at this level are fixable with targeted code changes.

50-69: Significant room for improvement. Multiple serious issues affecting real user experience. Keyboard navigation, form labels, or image accessibility likely have gaps. This is the range where users with disabilities will encounter real barriers.

Below 50: Major barriers present. Critical issues that block user access. Structural problems like missing landmarks, keyboard traps, or fundamentally inaccessible interactive components. These sites need prioritised remediation starting with the most severe issues.

Why scores from different tools don't match

If you scan the same site with Lighthouse, WAVE, and Lumi, you'll get different numbers. This isn't a bug - it's because each tool:

• Checks different rules (Lighthouse checks roughly 50; Lumi checks 350+)
• Weights issues differently
• Has different detection capabilities (DOM-only vs real browser testing)

A site can score 100 on Lighthouse and 45 on Lumi because Lumi's additional engines find issues Lighthouse doesn't check for.

The lesson: don't compare scores across tools. Use one tool consistently and track your trend over time.

What to do with your score

Don't aim for 100. A score of 100 means the automated tool found no issues. It does not mean your site is fully accessible. Remember: automation catches 30-50% of issues.

Do track the trend. Your score this month versus last month tells you whether things are getting better or worse. A consistent upward trend matters more than the absolute number.

Do fix critical and serious issues first. Three critical fixes will improve your score and your users' experience more than twenty minor fixes.

Do pair with manual testing. The score is a useful proxy, not a complete picture. Complement it with manual audit for full WCAG compliance.

Scan your site free at lumi.livana.io and see where you stand.