Understanding WCAG Severity Levels

A page can come back from an automated accessibility scan with forty findings, and it's tempting to treat that list as forty equally urgent problems. It isn't. Accessibility testing tools generally classify issues by impact — critical, serious, moderate, and minor — and that classification exists precisely so you can triage instead of drowning.

What each level actually means in practice

Critical issues typically block a whole category of user from completing a task at all — a form that can't be submitted without a mouse, an interactive element with no accessible name that a screen reader just announces as "button," content hidden behind a color-only distinction that conveys no information otherwise.

Serious issues create major friction without necessarily making something fully impossible — insufficient color contrast on body text, a focus indicator that's technically present but nearly invisible, images with alt text so vague it doesn't convey the image's actual content.

Moderate and minor issues are real and worth fixing, but they're refinements rather than blockers — a redundant title attribute duplicating visible text, a heading level skip that doesn't break navigation but isn't ideal structure.

Why this ordering matters for actually fixing things

If you work through findings alphabetically or in whatever order a report happens to list them, you'll often spend real engineering time on a minor issue while a critical one — something that fully locks a category of user out of a core task — sits untouched. Fixing by severity first means the people most affected get unblocked soonest, and it's also just a more defensible use of limited time.

Automated scanning finds a slice, not everything

It's worth being honest about what automated tools can and can't catch. They're excellent at structural, programmatically-detectable issues — contrast ratios, missing labels, invalid ARIA usage. They can't tell you whether your site's actual task flow makes sense to someone navigating by keyboard alone, or whether your error messages are genuinely clear when read aloud. Automated scanning is a strong first pass, not a substitute for occasionally testing with real assistive technology yourself.