Accessibility basics that also improve conversion

Accessibility gets presented to small businesses as compliance, which makes it feel like a cost with no return. That framing is wrong and it is why nothing gets done.

The honest version is simpler. The conditions that make a site hard for a person with a disability are the same conditions that make it hard for a fifty-five year old homeowner on a cracked phone screen in bright sun. You are not building for a small minority. You are building for the average visit.

A website is a piece of equipment. Equipment gets judged by output, not by appearance, and the output here is a call, a booking, or a form. Everything below is graded against that.

Which changes do the most work?

Contrast. Light grey text on white is the default look of a lot of modern templates and it is the most common accessibility failure on small business sites. The World Wide Web Consortium sets the AA threshold at a contrast ratio of 4.5 to 1 for normal text and 3 to 1 for large text, in success criterion 1.4.3 of WCAG 2.2. Mid grey text at hex 999999 on white measures 2.85 to 1 and fails. Darkening it to 767676 gives 4.54 to 1 and passes. Going to 595959 gives 7 to 1, which is comfortable rather than borderline. Any free contrast checker runs the same formula, so you can verify those three numbers yourself in a minute.

Text size. Body text set small looks tidy on a designer's monitor and forces a pinch-zoom on a phone. If a visitor has to zoom to read your hours, they will often just leave.

Tap targets. WCAG 2.2 added a specific number for this. Success criterion 2.5.8, at level AA, asks for targets of at least 24 by 24 CSS pixels, and the enhanced version at level AAA asks for 44 by 44. Two links stacked close together produce mis-taps, and a mis-tap that lands somebody on the wrong page usually ends the visit.

Form labels. Placeholder text inside a field disappears when you start typing. That is genuinely bad for screen readers and genuinely annoying for everyone, because people forget which field they are in and abandon the form.

Alt text. A short, factual description of what the image shows, not a keyword phrase. Screen readers read it, and it appears when an image fails to load.

Every item on that list makes the site easier for a customer who is in a hurry, which is all of them.

How does this actually affect conversion?

Through friction, not goodwill. Nobody calls you because your contrast ratio is good. Plenty of people fail to call because they could not read the number or could not hit the button.

There is a measured version of this. Google's mobile speed research with SOASTA, published on Google's own Think pages and updated in February 2018, trained a model on bounce and conversion data and reported that as the number of elements on a page goes from 400 to 6,000, the probability of conversion drops 95%. Element count is a proxy for clutter, and clutter is what the accessibility basics remove.

The clearest single example is the phone number. On a phone it needs to be text, and it needs to dial when tapped. A number baked into an image cannot be tapped, cannot be read aloud, and cannot be copied. This one item is missed constantly, and the click to call problem mobile businesses are still living with is the fastest revenue fix in this whole category.

What does the arithmetic look like on a phone number?

Is alt text a ranking factor?

No, and this is worth naming because alt text is routinely sold inside search packages.

Whitespark's Local Search Ranking Factors report, published 6 November 2025, had 47 local search experts score 187 factors. Keywords in image alt attributes scored 43. Presence of keywords in image metadata fields scored 30. For scale, the top factor in the same report, primary Google Business Profile category, scored 227. Alt text sits with the items those experts believe do essentially nothing for local rankings, alongside geo-tagged photos at 23, which is the lowest score on the list.

Grade that honestly: it is expert opinion, not a controlled test, and Whitespark says so. The direction is still unambiguous, and nothing supports selling alt text as a ranking service.

Write it anyway. It has a real job, it takes seconds, and the exercise is diagnostic. If you cannot describe what an image shows, the image contains no information. A stock photo of a handshake fails that test every time, which is one more argument for real photography on a small business website instead of purchased images.

What does the evidence actually say about form fields?

Less than the pitch does, and the gap is worth knowing before you pay for a form redesign.

Baymard Institute, which sells UX research and benchmarking, published checkout research in June 2024 finding that the average checkout flow runs 5.1 steps and 11.3 form fields, and that 17% of users have abandoned a purchase because checkout was too complicated. Baymard's load-bearing conclusion after more than a decade of testing is that step count matters far less than field count, and that overall performance falls as fields are added.

Here is the caveat that has to travel with it. That is e-commerce checkout research, not local lead-form research. It supports fewer fields as a direction. It does not support a specific claim about a plumber's quote form, and no study with a disclosed sample exists for contact or quote forms on local service websites. So when somebody tells you that cutting a form from eleven fields to four lifts conversions by a fixed percentage, ask for the study. That claim has no traceable primary source and is not a Baymard finding.

Measure your own form instead. Count the fields, delete every one you do not act on, label the rest, and compare submissions over a long window. That is not a controlled test, and you should say so out loud, but it is data about your business rather than somebody else's.

What about the compliance side?

There is legal exposure here, it varies by jurisdiction and by business type, and I am not the person to advise you on it. Ask your attorney if you want a real answer.

What I can tell you is what the public documents say. The Department of Justice published a web accessibility rule in 2024 adopting WCAG 2.1 Level AA as the technical standard, and that rule covers state and local government entities under Title II. It is not a rule for private businesses, and the Department's separate guidance for businesses under Title III does not name a technical standard. Anyone selling you a site on the claim that a specific federal standard binds your restaurant is going past what the documents say.

Be careful with overlay widgets, the products that promise to make any site accessible with one line of script. They do not fix the underlying markup, and a site with bad structure and an overlay on top is a site with bad structure.

Do the five basics properly instead. They are cheap, permanent, and they survive redesigns. The best moment is during a build, because you are already touching every page. Retrofitting costs more, which is one of the few genuinely good arguments for a rebuild when one is already justified. It is not, on its own, a reason to start one, and when a redesign is the wrong answer applies here too.

Where does this argument break down?

Three places, and the third one is the one people get wrong.

When the site is not the constraint. If almost nobody arrives, a more legible page reaches the same nobody. Accessibility multiplies attention. It does not create it.

When the fix is cosmetic rather than structural. Darkening text is a real fix. A widget that offers to darken text on request is not. The test is whether the improvement is in the page a screen reader receives, not in a control panel a visitor has to find first.

When you treat it as a project instead of a standard. The five basics decay. A new page ships with grey captions, a new form with placeholder-only fields, a new photo with no description. This belongs in the build brief and the publishing checklist, not in an annual audit.

Where else does this show up?

Outside your website, in the places customers actually meet you.

Photos on your Google listing get read by the same people with the same eyes, and text baked into a graphic is unreadable to anyone using assistive technology. If you publish updates through your profile, keep the words in the text field rather than inside an image. Whether Google Business Profile posts affect anything is a separate question, but if you are going to publish them, publish them readably.

It also shows up in whether your site still describes reality. Alt text written for services you stopped offering, a photo of a location you closed, a menu item removed two years ago. Those are accessibility problems and honesty problems at once, and they are among the signs a website was built for a business you no longer run.

What to do this week

Open your homepage on your phone, outside, in daylight. Try to read the body text without zooming. If you cannot, that is the first fix.

Tap your phone number. Confirm it dials. Then check whether the tap is recorded anywhere, because an untracked fix cannot be evaluated.

Fill out your own contact form, on a phone, with one field blank on purpose. Read the error message. If it does not name the field and the fix, that is the second job.

Run any free contrast checker against your text and background colors and get to 4.5 to 1 or better on body text. Darken until it passes.

Then pick your ten most important images and write one honest sentence of alt text for each. If you cannot describe what an image shows, replace the image.

Questions about your site's basics?

Email me at eric@seod.com with your website address. I will check three pages for contrast, text size, tappable phone number, and form labels, then send you the specific failures with what to change. It is a short list every time, and most of it is fixable in an afternoon by whoever maintains your site.

More on making a site usable in the web design and branding library.