Accessibility and SEO: the fixes that do both jobs
Accessibility is usually framed as compliance. A large part of the required work also improves search performance and conversion for everybody, and that part is worth doing first.
Accessibility gets budgeted as compliance, which puts it in the same queue as insurance renewals: necessary, unloved, and postponed until something forces it. That framing loses the part of the work that pays for itself twice, once in conversion and once in search.
The overlap is not a coincidence. A screen reader and a search crawler are doing the same job. Both consume your page without seeing it. Both depend entirely on the structure and the text you gave the markup. A page that is hard for one is usually hard for the other.
Where accessibility and SEO are the same work
Heading structure. Screen reader users navigate by jumping between headings, which is why a page needs one H1 and a sensible descent through H2 and H3. Search engines use the same structure to work out what a page is about and which passage answers which question. A page where headings were chosen for their font size, which is most pages built without a system, reads as a flat wall to both.
Descriptive link text. “Read more” tells a screen reader user nothing when they pull up a list of links on the page, which is a normal way to navigate. It tells a search engine nothing about the destination either. Link text is one of the clearest signals you can send about what a page you are linking to is about, and “click here” spends that signal on nothing.
Alternative text on images. Written for people who cannot see the image, and read by image search for exactly the same content. A diagram carrying real information with an empty alt attribute is invisible to both.
Semantic markup. A real <button> announces itself as a button, can be reached by
keyboard, and works without JavaScript. A styled <div> does none of that, and a crawler
following your conversion path hits the same wall.
Page speed and layout stability. Content that jumps while loading is disorienting for everyone and disabling for people with motor or cognitive impairments. It is also a ranking signal, measured directly.
None of these are compromises. Doing them properly serves both audiences with one piece of work.
Where it overlaps with conversion instead
A different set of fixes has no search effect at all and shows up in your conversion rate.
Contrast ratios exist so people with low vision can read text. They also decide whether anyone can read your call to action on a phone in daylight. The most common failure we find is a brand colour used for small text at a ratio around 4:1, which passes nobody’s eye test and fails the standard.
Form labels exist for screen readers. They also stop sighted users guessing what a field wants after the placeholder text vanishes on first keystroke. Placeholder-only forms measurably increase errors, and error rate on a form is conversion rate wearing a different name.
Visible focus states exist for keyboard users. They also tell anyone tabbing through a checkout where they are. Removing the focus ring because it looked untidy is one of the most expensive small decisions in web design.
Touch target sizing exists for people with limited dexterity. It also fixes the mis-taps that make mobile forms frustrating for everybody.
Roughly a third of a typical accessibility audit is work you would do anyway if you were optimising for conversion, and another third helps search. Sequence those two thirds first and the programme funds itself before you reach the parts that do not.
Where it stops overlapping, and how to argue for it there
Some of the work genuinely is compliance with no commercial upside, and pretending otherwise damages your case with the people who approve budgets.
Full keyboard operation of a complex interactive component. Correct ARIA on a custom widget. Captions and transcripts for video. Reduced-motion alternatives. These take real engineering time and will not move a conversion rate.
Argue those on the correct grounds. There is legal exposure, and it is rising in most jurisdictions. There is the tender question, because public sector and enterprise procurement increasingly asks for a conformance statement and a missing one loses the deal before anyone reads your pitch. And there is the plain fact that excluding people from your site is a choice somebody made, usually without noticing.
Do not stretch the commercial argument to cover this third. Make the commercial argument for the two thirds where it is true, ship those, and let the demonstrated value buy you the room to finish.
A sequence that works
Order the work by cost against combined benefit, not by the order the audit tool printed.
- Contrast. Cheapest fix in the list, affects every page, and it is measurable in both an audit and a conversion test. Usually a handful of token changes if your design system has tokens, and a long afternoon if it does not.
- Focus states. One stylesheet change. Restores keyboard navigation across the whole site.
- Form labels and error messages. Contained work, direct conversion effect, and it takes your most commercially important page out of the failure list.
- Heading structure. Fixes accessibility and gives search engines a real outline. Best done as part of a template pass rather than page by page, so it stays fixed.
- Link text and alt attributes. Content work rather than engineering, which means it can run in parallel with everything above.
- Keyboard paths through interactive components. The expensive part. Start with the ones on the conversion path.
The first three are usually a week. They will move your audit score more than anything else on the list, and they are the ones you can point at when someone asks what the budget bought.
Test it with the tools your visitors use
Automated checkers find perhaps a third of real issues. They are worth running in your build so regressions fail loudly, and they will not tell you whether your site can actually be used.
Three manual checks catch most of what the tools miss, and none needs a specialist:
Unplug your mouse and complete your main conversion flow using only the keyboard. If you get stuck or lose track of where you are, so does everyone who navigates this way.
Turn on the screen reader already installed on your machine and listen to one page. It is uncomfortable the first time and it is the fastest way to hear that your headings are decorative and your links all say “read more”.
Zoom the page to 200 percent. Text should reflow rather than requiring horizontal scrolling. This one catches layout assumptions no automated tool will report.
Every one of those three also surfaces problems that have nothing to do with disability and everything to do with whether your site works. That is the argument, and it holds without anyone needing to be persuaded that compliance is interesting.
