|
|
Why we’re writing about this
We build and manage websites for dental and orthodontic practices, so when the National Orthodontic Registry published a crawl of 1,683 orthodontic practice websites last month, we read every line. It confirmed what we see on nearly every audit we run for a new client: the practice doesn’t know what its website is doing, and the website is quietly making marketing, privacy, and AI-visibility decisions that nobody in the office signed off on.
The headline numbers: 22.5% of orthodontic sites still run Google Analytics code that stopped collecting data in July 2023. 24.5% of appointment pages carry a Meta Pixel. And whether a site can be read by ChatGPT, Claude, or Perplexity is decided almost entirely by the website platform’s default settings, not by the doctor.
This post walks through the findings, what they mean for your practice’s marketing, and the audit we now run on every orthodontic site we take over.
About the study
The State of Orthodontic Practice Websites 2026 is a technical crawl, not a survey. On September 24, 2026, the registry’s crawler visited 1,683 orthodontic practice websites and, for each one, read robots.txt, the homepage, llms.txt, the sitemap, and up to six appointment, contact, and treatment pages. It then fingerprinted the platform and matched 91 vendor signatures for tracking, booking, review, financing, and chat tools.
Two methodology choices make this more credible than the usual vendor-sponsored “state of” report:
• It downloaded 1,611 Google Tag Manager containers and read the tags inside. Most tracking on dental sites lives in GTM, and a page-source scan alone misses it. The study found, for example, that 80 of 309 CallRail installs and 68 of 87 PatientPrism installs were visible only inside Tag Manager.
• It validated against a real browser on 30 random sites, with 25 to 28 of 30 agreeing on the major tools.
The registry is funded partly by paid placement, but every practice in its directory was measured the same way, and the authors state that no company named in the report paid for the research. The aggregate data is downloadable under a CC BY 4.0 license. The authors are also upfront that most figures are minimums, since tracking that fires only after consent or a form submission is invisible from outside.
The findings that matter to your marketing
Platform defaults decide your AI visibility. 29.8% of orthodontic sites serve an llms.txt file (a plain-text summary written for AI models), but the number is almost pure platform: 97.1% of Duda sites and 92.0% of Wix sites have one, versus 2.6% of Sesame and 2.4% of Squarespace sites. On WordPress it depends on the SEO plugin: 82.1% with All in One SEO, 44.2% with Rank Math, 25.1% with Yoast. On the blocking side, only 2.2% of sites block a major AI crawler, but 45 of the 69 sites blocking GPTBot are ProSites builds, because ProSites ships a robots.txt that allows named search engines and blocks everything else. Six sites in the sample block Googlebot outright.
One in four consultation pages is sending data to Meta. Of 1,355 sites with an appointment or consultation page, 24.5% load Meta Pixel code there. In a browser test of 38 such pages, the pixel fired on 30. Only 9.2% of sites have a cookie-consent tool and 1.9% use analytics marketed as HIPAA-compliant. The study didn’t submit forms, so post-submission behavior wasn’t tested, and it explicitly does not assess legal compliance.
Analytics is unmanaged on a third of sites. 22.5% still run Universal Analytics, which Google retired on July 1, 2023. 14.9% have no Google Analytics at all. Meanwhile 40.1% carry Google Ads conversion tracking and 48.8% use Tag Manager, so plenty of practices are paying for ads while their measurement is broken or missing.
The ortho tech stack is fragmented. 27.3% use a third-party booking or lead-capture tool, and no vendor exceeds 7% share (Jotform 6.5%, SmileSnap 4.7%, NexHealth 4.2%). Patient-portal tools reach 41.7%, led by Sesame at 14.0% and Ortho2 at 9.4%. Review widgets are on 18.4% of sites, call tracking on 16.2%, financing embeds on 7.0%, and AI receptionists on just 2.0%. 31.0% load an accessibility overlay, mostly UserWay.
WordPress dominates; niche dental builders hold a sixth. 59.6% of sites run WordPress. Dental- and orthodontic-specific platforms hold 17.3%, led by Sesame (10.5%) and Rooster Grin (3.3%). Among agencies credited in site footers, Sesame Communications (172), Neon Canvas (97), and HIP Agency (58) lead.
Finding
Share of orthodontic sites
Serve an llms.txt with real content
29.8%
Block a major AI crawler
2.2%
Local-business schema markup
54.8%
Meta Pixel on appointment page
24.5%
Cookie-consent tool
9.2%
Still run Universal Analytics
22.5%
No Google Analytics
14.9%
Google Ads conversion tracking
40.1%
Google Tag Manager
48.8%
Third-party booking tool
27.3%
Accessibility overlay
31.0%
Built on WordPress
59.6%
Source: National Orthodontic Registry, September 2026
Region is close to irrelevant. The four U.S. regions land within a few points of each other on every measure; what platform a site is built on explains far more than what state it’s in.
What this means for your practice
You’re probably buying ads with broken measurement. Four in ten orthodontic sites run Google Ads conversion tracking, yet one in five still runs analytics that has collected nothing since 2023 and one in seven has no analytics at all. In our experience those groups overlap heavily. If your agency reports “conversions” but your GA property has been dark for three years, nobody can tell you what a new-patient consult actually costs. Measurement should be the first line item on any marketing engagement, not an afterthought.
Your consult page is a privacy and reputation risk you didn’t choose. Meta retargeting is a legitimate tactic on a treatment page. On the page where a prospective patient types in their name, phone, and reason for visiting, it is a different conversation. The federal guidance on healthcare web tracking was partly vacated in American Hospital Association v. Becerra in June 2024, so the legal picture is unsettled. But patients increasingly know what a pixel is, and a practice that can’t explain what its own consult page sends to Facebook is exposed regardless of how the law shakes out. With only 9.2% of sites running consent tooling, most practices have no defensible answer today.
AI search is the next referral channel, and your vendor already decided your position in it. A growing share of “orthodontist near me” queries now go through ChatGPT, Perplexity, Google’s AI Overviews, and Claude. Those tools need two things: permission to crawl you, and structured data that tells them who you are and where. The study found 45% of sites lack a local-business schema, and the crawl-permission decision is being made by Wix, ProSites, or a WordPress plugin default. A ProSites client, for instance, is almost certainly blocking GPTBot and has no idea. We’d treat llms.txt as a cheap hedge rather than a proven lever, since no major AI provider has committed to reading it, but schema markup and crawler access are table stakes.
The vendor you chose is a marketing decision. This is the study’s real lesson. A practice on Wix has an llms.txt it never asked for. A practice on ProSites blocks AI crawlers it never heard of. A practice on Sesame almost certainly doesn’t serve llms.txt but very likely has the Sesame portal. None of that was decided by the doctor. When you evaluate a website vendor or agency, ask what their defaults are for crawler access, pixels on form pages, consent tooling, and analytics hygiene. If they can’t answer in a sentence, that tells you something.
There is a competitive opening. Because the numbers are driven by defaults rather than strategy, a practice that handles the basics deliberately is ahead of roughly 70% of its local competitors on AI readiness and ahead of 90% on privacy posture. That is unusually cheap differentiation.
The audit we run on every orthodontic site
This is the checklist we now work through on day one with a new practice. Most of it takes an afternoon; all of it should be repeated yearly, which is also how often the registry plans to rerun its crawl.
1. Tag inventory. Export every tag from Google Tag Manager and every script in the page source. Name the owner and purpose of each. Delete anything without one, starting with Universal Analytics.
2. Analytics baseline. Confirm GA4 is installed, receiving data, and that consult-request submissions are tracked as conversions. If Google Ads is running, verify its conversion tags fire on a real submission, not just a page view.
3. Form-page pixel policy. Decide, in writing, which pixels are allowed on appointment, contact, and patient-form pages. Our default recommendation is none on form pages; retargeting belongs on treatment and homepage content.
4. Consent and privacy policy. Install a consent tool and rewrite the privacy policy to name the actual tools in use. Consider HIPAA-oriented analytics if the practice runs patient forms through the site.
5. robots.txt review. Check that the site isn’t blocking GPTBot, ClaudeBot, PerplexityBot, Google-Extended, or Googlebot because of a vendor template. Decide crawler access on purpose.
6. Structured data. Add or fix LocalBusiness and Dentist/MedicalBusiness schema with name, address, phone, hours, and services. Verify it in Google’s Rich Results test.
7. llms.txt. Publish a short, accurate one. Low cost, uncertain payoff, no downside.
8. Booking and tool sprawl. List every booking, review, financing, chat, and call-tracking embed. Each one is a script, a vendor relationship, and a data flow; keep the ones that produce measurable consults.
9. Calendar it. Put the next review twelve months out.
The practices in this study aren’t careless. They hired a vendor, got a website, and moved on to running an orthodontic office, which is exactly what they should be doing. But the study makes clear that “the vendor handles it” has quietly turned into “the vendor’s defaults decide it.” Someone should be choosing on your behalf, with your sign-off. If that isn’t happening today, start with the tag inventory. It’s the step that surfaces everything else.
Study: The State of Orthodontic Practice Websites 2026, National Orthodontic Registry, crawled September 24, 2026. All figures above are from that report.
|
Locked
