We measured 1,756 Australian dental practice websites

Everyone in dental marketing has an opinion about what practice websites get wrong. Nobody publishes the measurement. So we took every Australian dental practice we could resolve from public sources, opened each site in a real mobile browser, and recorded what was actually there. This page is the result, measured in August 2026.

What we measured, and what fell out along the way

The denominators matter more than the percentages, so here is where the number shrinks at each step. Every figure further down is a share of the last row, not of the first.

152 sites did not load at all — timeouts, refused connections, TLS failures or server errors, retried and still dead. That is about 7% of every dental website we tried, and it is the single starkest number here: for one practice in fourteen, a patient who types the name into Google and clicks through gets nothing.

Google cannot read most of them

• 1,071 of 1,756 (61%) publish no structured data identifying them as a dental practice. Their address, opening hours and services are not machine-readable, so the Google Business Profile and the website are, as far as Google is concerned, two unrelated things.
• 418 (24%) have fewer than three service pages. There is almost nothing for Google to rank against a treatment search — someone searching "wisdom tooth removal" plus a suburb is matched against a homepage.
• The median homepage carries 547 words, and 801 sites are under 500.

Structured data is not a ranking trick and we are not claiming it is one. It is the machine-readable version of the facts already on the page, and its absence is why a practice can be first on the map and invisible in a treatment search.

Almost nobody has enough traffic for Google to measure them

Google collects Core Web Vitals from real Chrome users and publishes them in the Chrome UX Report. A site with too few visitors does not appear in it at all — and that absence is itself a finding.

• 1,785 of 1,951 (91%) have no Core Web Vitals field data whatsoever.
• Of the 166 that do have real-user data, 90 (54%) fail Core Web Vitals on mobile.
• Median time to show the main content, measured in the lab: 2.4s — inside Google’s 2.5s threshold. The median hides the tail: the slowest quarter sit above 3.2s.
• On the median site, 8% of real visits wait more than four seconds before anything appears.

There is often no way to get in touch

• 450 (26%) have no enquiry form of any kind.
• 353 (20%) have no form, no chat and no booking link. Outside opening hours the only route in is a phone nobody is answering.
• 220 (13%) do not make the phone number tappable — it is text on a phone screen, not a link.
• 658 (37%) do not show a phone number in the first screen on a phone at all.

One in five practices, then, is uncontactable after hours by anything except a phone call to a closed practice. That is the finding we expected least and the one we would fix first.

On a phone

• 281 (16%) render content past the right edge at 390px — text or buttons cut off on a common phone width.
• The median site has 24 tap targets under 44px.
• 11 (under 1%) were not served over HTTPS, so Chrome marks them Not Secure.

Who is already being paid to look after them

928 (53%) carry advertising pixels, which means they are buying traffic. 235 run a paid review platform, 137 a paid chat widget and 242 a paid booking system. Read alongside the findings above, that is the uncomfortable part: a majority are paying to send people to a site Google cannot read and a fifth of them cannot be contacted out of hours.

Reviews are not distributed the way people assume

Across the 1,074 practices we hold Google review counts for, the median is 60 reviews at a median rating of 4.8. The average is not the story. The 25th percentile is 17 reviews and the 75th is 177, with a maximum of 4,197 — in most suburbs one practice has taken the reviews and everyone else divides what is left.

A median rating of 4.8 across an entire profession also means rating is close to useless as a differentiator. Volume and recency are what separate two practices that both sit at 4.8.

What this adds up to

The common failure is not that these websites are ugly. Most are fine to look at. It is that the website and the Google listing are two unconnected things, the site is invisible to the measurement Google actually uses, and there is no way to make contact at the hour people are looking. Those three are among the cheapest things to fix, and they are the three nobody has been sold.

How many dental practices are there in Australia?

We resolved 4,878 distinct practices from five public sources and judged 4,331 of them addressable — independent, dental, and contactable. That is a floor rather than a census: a practice with no public listing anywhere would not appear in it.

Does my practice website have structured data?

Open your homepage, view source, and search for "application/ld+json". If nothing comes back, you have none. If something does, paste the page URL into Google’s Rich Results Test to see whether it identifies you as a dental practice.

Why does having no Core Web Vitals data matter?

It does not mean your site is slow. It means Google has not seen enough real Chrome visitors on it to publish a figure — a traffic signal rather than a speed one. It also means you cannot prove your site is fast when someone claims otherwise.

Is an enquiry form without a consent checkbox illegal?

Not automatically, but dental enquiry forms usually collect a health concern in the message, and that is health information under the Privacy Act. Collecting it needs consent and a privacy notice that says what happens to it. A checkbox and a linked policy is the ordinary way to show that.

Can I get my own practice’s figures?

Yes. We hold a per-practice version of everything above and we will send you yours, dated, free, with no obligation. We do not publish or sell an individual practice’s findings to anybody else.

Method

Measured in August 2026. Practices were resolved from five public sources and de-duplicated on a normalised name, domain and phone number. Each website was opened in headless Chromium at 390×844 in en-AU on Australia/Sydney time, and every finding was read out of the rendered page rather than the HTML source, because a site that builds its content in JavaScript looks empty in source and fine to a visitor.

• Core Web Vitals field data comes from the Chrome UX Report; lab timings come from PageSpeed Insights on mobile.
• Review counts and ratings come from Google Maps records. We store the count and rating we measured and the date we measured them, and no review text or reviewer name is stored anywhere.
• Structured data is counted only when it identifies the business as a dental or medical practice. Generic site-level markup does not count.
• "No enquiry form" means no form, of any kind, anywhere on the pages we crawled. A form on a page we did not reach would be counted as absent, which makes that figure a maximum rather than an exact count.
• Percentages are of the row above them in the coverage table, always stated. Nothing here is a share of 4,878 unless it says so.