Among the many ancillary arguments you tend to hear about why you should care about accessibility is the pitch: “Accessibility improves your SEO!” It’s the sweetener accessibility advocates add add when they suspect compliance alone won’t close the deal. And like most such arguments, it contains just enough truth to survive scrutiny – as long as nobody actually scrutinizes it.

We did. We combined the results of 20 real-world accessibility audits spanning 286 distinct manual checks each and over 100,000 automated test failures from leading accessibility tools. Then we asked a simple question: of the accessibility problems that actually occur in the wild, how many would a search engine ever notice?

The answer: not many, especially among accessibility issues that matter most.

Yes, there’s some overlap with SEO and accessibility

The “accessibility helps SEO” claim isn’t a fabrication. A handful of practices genuinely serve both:

Page titles, heading structure, image alternative text, descriptive link text, and links built with real <a href> markup all matter to assistive technology users and to search engines. In our data, the single most frequent automated failure across all 20 audits was the use of non-descriptive, ambiguous link text and links to different destinations sharing identical text. That’s a legitimate accessibility failure and legitimately bad anchor text which will impact SEO. Similarly, we found more than 3,000 “links” that weren’t real links at all. They lack a valid hypertext reference and are, instead, just JavaScript stapled to a <div> that are meant to simulate a button. A screen reader can’t find them, and neither can Googlebot.

So we counted generously. We classified every check with any plausible SEO connection as overlapping: all checks relating to titles, all heading checks, every alt-text check, all link text quality, etc. We even threw in marginal items like the lang attribute and mobile zoom behavior.

The result: 13 of 127 automated checks have direct SEO relevance. They account for about 27% of automated failure instances. In the manual audit data, the picture is worse: fewer than 16% of all failures observed across 20 audits touch anything a search engine cares about. The other 83% are invisible to SEO.

The most frequent and most severe failures mean nothing to Google

The single most consistently failed check in our entire dataset concerns dynamic content updates. These are things like status messages, live region announcements, and content that changes without notifying assistive technology. These checks failed in every single audit where they were tested. A perfect failure rate. In such cases, real end users are impacted but it has no impact for SEO. There is no ranking factor for “your form errors are announced to assistive technology.”

In 2nd place were four ARIA checks with 100% failure rates. ARIA failures broadly run at a 54% fail rate across audits for things like invalid properties, broken references, misused roles, and focusable elements hidden from assistive technology. ARIA exists exclusively to communicate with assistive technology. Search engines have publicly and repeatedly confirmed they don’t use it for ranking.

Color contrast failed in roughly 79% of the audits where it was evaluated, and automated tools logged more than 7,000 contrast failures across these properties. Contrast is one of the most common reasons real users with low vision can’t use a site. Googlebot does not perceive color.

Keyboard accessibility failed in nearly 60% of audits. Googlebot does not use a keyboard. It does not tab through your checkout flow. Inoperable controls make your site literally unusable for a motor-impaired customer registers exactly nowhere in your search performance.

Forms were the single largest category in our manual testing, with failures spanning missing labels, broken groupings, and inaccessible error handling. Whether a screen reader user can determine what your a field is asking for has no bearing on whether your page ranks for anything.

When we looked at the twenty most frequently failed manual checks across all audits, seventeen had no SEO relevance whatsoever. The three exceptions: heading hierarchy, alt text on actionable graphics, descriptive link text are real, and worth fixing. But they’re outnumbered nearly six to one by failures in dynamic content, ARIA, contrast, visual focus indication, form labeling, and keyboard operation.

Why the disciplines diverge: search engines model one user

Accessibility advocates love to say “Googlebot is blind”. And, while that’s true for things like good page titles, headings, and text alternatives, the biggest difference is that it performs no interactions and runs no assistive technology. It merely parses your content once, top to bottom. SEO optimizes for that one synthetic, non-interactive reader.

Accessibility is the opposite problem: it’s about every other kind of user, such as the screen reader user who needs your ARIA to be valid, the keyboard-only user who needs focus order to match visual order, the low-vision user who needs 4.5:1 contrastm and the user with a cognitive disability who needs your form errors to be clear and your content not to move unexpectedly. None of these users are modeled by a crawler, so none of their barriers show up as ranking signals.

This is also why the “accessibility overlaps with SEO” framing has cause and effect backwards. The overlap exists because both disciplines inherited a sliver of the same foundation: semantic HTML. Titles, headings, alt text, and real links are good markup hygiene. SEO uses that sliver because it’s machine-readable. Accessibility starts there and then extends into interaction, perception, and comprehension. This is territory no search bot enters.

Sometimes SEO advice is actively hostile to accessibility

In a few places, common SEO tactics directly conflict with accessibility best practice, such as text alternatives. SEO guidance frequently says to put descriptive, keyword-rich alt text on every image. Accessibility says decorative images should be programmatically ignored so assistive technology skips them. In our audits, improperly implemented decorative images failed in 72% of the websites we tested. A site that follows keyword-everywhere alt advice forces screen reader users to sit through announcements of every flourish, divider, and stock photo on the page. The SEO tactic creates the accessibility failure.

Another divergence is with links that open in new windows. Some SEO practitioners still like target="_blank" to keep visitors “on site” for dwell-time metrics. Automated testing tools logged over 5,000 failures for exactly this pattern, because unannounced new windows disorient screen reader users and break the back button that many users with cognitive or motor disabilities rely on.

The judgment gap: SEO tools can’t even see the worst problems

One more finding from our analysis deserves attention. When we compared automated coverage against manual results, we found 60 checks that fail in real audits but have no automated test at all. 13 of those checks had manual failure rates above 75%. ARIA as a category has only 28% automated coverage. Dynamic content: 20%. Keyboard: 35%.

The most damaging accessibility failures require human judgment to detect, such as focus management, the clarity of error messages, or proper implementation of widget patterns. No SEO crawler, no Lighthouse score, no site-audit SaaS will ever surface them, because they’re not detectable by the kind of inspection those tools perform.

What to do with this

None of this means the overlap work isn’t worth doing. Fix your link text, because bad link text fails in two-thirds of our audits and it will help both audiences. Fix your headings, your titles, your alt text. These are the cheapest, most automatable quarter of the job. But buy accessibility for what it actually is. Across our 20 audits, failure rates ranged from 10% to 100% of tested checks – including organizations you’d expect to know better. The failures that expose you to legal risk and that actually exclude human beings represent 75–83% of the problem space that search engines will never see, never reward, and never penalize.

If a vendor leads with the SEO benefits of accessibility, they’re selling you the smallest part of the work. The rest of it, such as whether a blind user can buy your product or a keyboard user can submit your form has exactly one return on investment: people can use your website. That is why you should make your site accessible.