Conversion · 9 min read

Dental Websites Do Not Have an Alt Text Problem. They Have a Booking Form Problem.

  • Home
  • Blog
  • Dental Websites Do Not Have an Alt Text Problem. They Have a Booking Form Problem.

We rendered 20 randomly drawn Canadian dental websites in a real browser and checked what a screen reader would find. The alt text is mostly fine: 15 of 20 sites have no image missing alt at all, and 94% of every missing-alt image in the sample comes from just two sites. The failure is in the form. 28.5% of the 123 form fields we measured have no accessible name, 8 of the 13 sites with a form have at least one, and on five of them not a single field in the form is labelled. That is the page that takes bookings.

By ScaleUp Dental·Original research, November 2026·Published 2026-11-13

Ontario’s accessibility regulation requires WCAG 2.0 Level AA for websites, and the threshold for that requirement is organisations with 50 or more employees, plus the public sector. A typical dental practice has nowhere near 50 employees, so for most Canadian clinics the website standard does not apply. Other parts of the accessibility framework, such as accessible customer service, do.

We are stating that from secondary sources rather than the regulation itself, and a practice wanting a legal answer should read O. Reg. 191/11 or ask a lawyer rather than an SEO agency. But the direction matters for how to read everything below: nobody is going to fine a seven-person clinic for its website. This is not a compliance article. It is an article about whether a patient who cannot see your form can book an appointment.

The alt text panic is mostly misplaced

The standard accessibility pitch to a small business leads with missing image descriptions. Across 782 images on these 20 sites, 130 have no alt attribute — 16.6%, which sounds like it supports the pitch until you see where they are. 122 of those 130 are on two websites. The other 18 sites account for 8 between them, and 15 of the 20 have none at all.

Something else worth saying plainly, because we got it wrong ourselves in an early audit: an empty alt attribute is not a failure. For a decorative image it is the correct markup, and it tells a screen reader to skip the image rather than announce a filename. 16.8% of the images in this sample carry a deliberate empty alt. An audit that counts those as errors is manufacturing a problem, and we have seen that number quoted at clinics as if it were one.

  • 15 of 20 sites have no image missing an alt attribute.
  • 94% of all missing-alt images in the sample sit on 2 sites.
  • 16.8% of images use alt="" deliberately, which is correct for decorative images.
  • 19 of 20 sites have exactly one h1.

The form is where it actually breaks

13 of the 20 sites have a form on the homepage. Across them, 35 of 123 fields — 28.5% — have no accessible name: no label, no aria-label, no wrapping label, nothing a screen reader can read out. The field is announced as "edit text, blank".

On five of those thirteen sites, that is true of every field in the form. A person using a screen reader reaches the booking form, hears a series of unlabelled boxes, and has no way to know which one wants the phone number.

This is the part that costs something. Our earlier study of these forms found the median asks for five or six fields and that the form is often the only path besides the phone. A form nobody can fill is not an accessibility footnote; it is the conversion path, broken for a specific group of patients.

Link text is a smaller version of the same problem. 4.0% of 2,708 links have no accessible name and 3.0% are "click here" or "read more" — half the sites have at least one. One site has 36 of them.

What we would fix, in order

Label every form field first. It is the smallest change on this list and the only one on the booking path. A visible label associated with the input is better than a placeholder, because placeholders disappear the moment someone types and are not reliably announced.

Then the two sites with 81 and 41 missing alt attributes, which is a template problem rather than a content problem and is usually one fix repeated.

Then the link text, which also helps the ordinary visitor scanning the page.

Heading order is last. 65% of the sample skips a level somewhere, which is the most common failure here and the least consequential — a screen reader user navigating by heading still gets where they are going.

  • Form fields: a real label, associated with the input. Not a placeholder.
  • Images: alt that describes the image, or alt="" if it is decoration.
  • Links: text that says where it goes, instead of "read more".
  • Headings: in order, once the above are done.

Why we rendered the pages instead of reading the HTML

We ran a static pass over the served HTML of all 194 reachable clinics first, because that is fast and free. It is also what the cheap accessibility scanners do, and the numbers it produced were wrong.

Two of the three sites we spot-checked were false positives. One was reported as having all seven form fields unlabelled; rendered, all seven carry a label correctly associated with the input. The labels are added by script after the page loads, which a parser reading the served HTML never sees. A second bug in our own checker counted collapsed menu links as having no name, because the method we used to read link text ignores anything hidden by CSS — one site went from 40 nameless links to zero once we fixed it.

So we discarded the 194-site figures and rendered a 20-site sample instead. That is the second time this year a static audit has given us a confidently wrong answer about these websites, and the lesson is the same: if a report lists accessibility errors on your site, open the page before you pay anyone to fix them.

Limits

Twenty sites, randomly drawn, rendered and checked individually. A different draw moves the percentages, and with two sites carrying most of the missing alt attributes the image figures in particular are sensitive to which sites come up.

We checked homepages only, and only the things a machine can judge: accessible names, alt attributes, heading order. We did not test colour contrast, keyboard traps, focus order or screen reader output, which need a person and are where the harder problems live.

Passing these five checks does not make a site accessible. Failing them does make it harder to use. Nothing here is a legal opinion.

Accessibility checks on 20 Canadian dental homepages, rendered in a browser (ScaleUp Dental, November 2026).

CheckResult
Images missing an alt attribute130 of 782 (16.6%) — 94% on two sites
Sites with no missing alt at all15 of 20
Images using alt="" correctly131 of 782 (16.8%)
Form fields with no accessible name35 of 123 (28.5%)
Sites where no form field is labelled5 of 13 with a form
Links with no accessible name109 of 2,708 (4.0%)
Links reading "click here" or similar80 of 2,708 (3.0%)
Sites skipping a heading level13 of 20 (65%)
Sites passing all five checks5 of 20

FAQ

Do Canadian dental websites have to be accessible by law?

Ontario’s WCAG 2.0 Level AA website requirement applies to organisations with 50 or more employees and to the public sector, so most dental practices fall below the threshold for that specific rule. Other accessibility obligations, such as accessible customer service, still apply. For a legal answer, read O. Reg. 191/11 or ask a lawyer — not an SEO agency.

Is empty alt text an accessibility error?

No. For a decorative image, alt="" is the correct markup — it tells a screen reader to skip the image instead of reading out a filename. 16.8% of the images in our sample use it deliberately. We counted it as a failure in an early audit of our own and were wrong; any report that flags it is inflating its error count.

What is the most common real accessibility problem on dental sites?

Unlabelled form fields. 28.5% of the form fields we measured have no accessible name, and on five of the thirteen sites with a form, not one field in it is labelled. A screen reader announces those as "edit text, blank". It matters more than the other failures because the form is the booking path.

Can I trust an automated accessibility report?

Only after opening the page. Our static pass over served HTML reported sites as having every form field unlabelled when, rendered, every field was correctly labelled — the labels are added by script. Two of three spot-checks were false positives, so we threw the static results away. Cheap scanners read HTML the same way.

What should a dental practice fix first?

Form labels. It is the smallest change, it is the only one on the booking path, and a visible label tied to the input beats a placeholder — placeholders vanish when someone starts typing. Images next, then link text, then heading order, which is the most common failure in our sample and the least consequential.

Grade your own clinic

Download the free 40-Point Dental Marketing Checklist or request a Free Dental Growth Audit.

Get the ChecklistFree Growth Audit
Free Consultation

Request a Free Dental Growth Audit

We help dental clinics turn Google traffic into tracked patient appointments.

ScaleUpDental
Services
Contact
Monday – Friday, 9:00 AM – 6:00 PM ET
Toronto, Ontario, Canada
[email protected]
Book a Free Consultation