The Therapist Recognition Index: 133 Therapist Websites Checked
By Dr. Nicholas Gehle, Psy.D. · Pages captured
This is the second edition. It uses a new method, so its numbers can't be compared with the first edition (July 2026), which stays published unchanged.
therapist websites that publish structured data for search engines and AI assistants don't name a clinician with a role or credential in it.
Most had the technical part in place. Almost none named the clinician.
Structured data (JSON-LD) is information a website publishes in its code for software to read: the practice's name, what kind of business it is, where it is, and who works there. It isn't part of the visible page. Of the 133 practice pages we checked, 109 published structured data we could read. Only 5 of those named a person along with a role or credential, such as "Jane Smith, Licensed Professional Counselor". Some of the other 104 listed a name or a business, but none named a clinician with a role or credential. We didn't test whether any search engine or AI assistant actually uses this information.
The headline findings
94 of the 100 practices whose pages show a credential don't name a clinician with a role or credential in their structured data. A credential such as LPC, LCSW or PhD appeared in the page text, title or search description, but the information published for software didn't name a clinician with a role or credential. Of the other 6, 5 did and 1 couldn't be checked. (This compares two checks per site; we didn't match the individual people.)
Only 19 of 132 declared a health or professional practice type in their structured data. Among the 109 pages with structured data, 90 didn't use one of the practice types our check recognizes (for example, "MedicalBusiness" or "ProfessionalService").
The basics are almost universal. All 133 had a page title, 133 had a contact route (a phone, email, booking or contact link), and 130 ended on a secure https address. The gap isn't the basics. It's whether the site says who the practice is and who provides the care.
On 8 sites, a contact detail didn't match its link. We found 9 cases where the email address or phone number written in a link differed from where the link actually goes. We didn't test whether calls or emails fail, or how visible each link is.
What we checked
Each page got 11 or 12 checks: ten for every page, plus one or two location checks depending on whether the practice looks local, online, or both (82 pages got 11 checks, 51 got 12). Each check comes back as found, not found by our scan, couldn't be assessed, or a confirmed "no" (used only for 3 pages that ended on a plain http address). "Not found by our scan" means our reader didn't establish it, not that the site definitely lacks it.
What a visitor's page contains
| Check | Found |
|---|---|
| Page title | 133 of 133 |
| A contact route (phone, email, booking or contact link) | 133 of 133 |
| Ends on a secure https address | 130 of 133 |
| Search description | 104 of 131 |
| A credential (in the page text, title or search description) | 100 of 133 |
| Exactly one main heading (15 couldn't be assessed) | 67 of 118 |
| City or state in the title, site header or main headings (local practices; 2 couldn't be assessed; menu links aren't counted) | 54 of 83 |
| Where online sessions are offered (online practices; 3 couldn't be assessed) | 55 of 75 |
What the structured data says
| Check | Found |
|---|---|
| Readable structured data present | 109 of 132 |
| An address or service area in it | 42 of 132 |
| A health or professional practice type | 19 of 132 |
| A named clinician with a role or credential | 5 of 132 |
Counts exclude checks that couldn't be assessed (one page for each structured-data row). Location checks: 34 pages looked local, 27 online, 51 both; 21 couldn't be classified and got an "unknown" location check instead.
How to read these numbers
These are counts of what we found, not grades. A missing check doesn't make a practice worse at therapy, and a found check isn't a guarantee of anything. Structured data being present doesn't prove that any search engine or AI assistant reads it or describes a practice correctly. It's the information a practice publishes for software, and in this group of practices it rarely named the person providing the care.
The practical takeaway is small and specific: if your site publishes structured data, it can name you, your role or credential, what kind of practice you are, and where you work.
Method
- Updated September 26, 2026: the reader now recognizes more of the ways practices describe online therapy (for example “in-person and online sessions”, “telemedicine”, “remotely across Tennessee”), and it only counts a compact such as PSYPACT as an online area when the practice says it offers it. The location figures were recomputed from the same saved pages; 11 pages changed. No other figure moved.
- Sample: the same 181-website cohort as the first edition: private-practice therapist websites drawn from public listings in 19 cities across 19 US states, labeled solo or group. One URL per practice, usually the homepage (2 of the 133 checked pages were a contact page and a clinician profile page). A multi-state descriptive sample, not a representative estimate for all US therapist websites.
- What happened to the 181: 133 were checked. 13 were refused by our reader: 5 had too little readable text, 4 gave conflicting readings, 3 had almost no content, and 1 was a placeholder page. Except for the placeholder, these are limits of our reader, not problems we can attribute to the site. 35 couldn't be captured: 26 connections failed, 6 pages were "not found", 2 refused access, and 1 was too large. Refused and uncaptured pages are reported as such, not counted as failures.
- Capture: every URL was requested on September 25, 2026. The 146 pages that returned HTML were saved and checked from the saved copy; the page reached after redirects is what was checked.
- Instrument: the Recognition Score's source checks. They are deterministic and rule-based, with no AI in the check path. They read selected parts of the page source: text in paragraphs, lists, headings and short text elements (up to a size limit), the page title and search description, links, and structured data in JSON-LD. They don't interpret images (the heading check can read image alt text), content drawn only by JavaScript, or other pages on the site. Text in other places can be missed.
- Independent reproduction: a separately written program re-ran all 181 saved records through the same checks and matched every count on this page. A manual review of the page source confirmed all 5 "named clinician" results and all 9 contact mismatches. The credential detections it found missing (credentials in smaller headings and short text elements, and several credential abbreviations) were fixed before these numbers were produced.
- What we did not measure: search rankings, traffic, referrals, bookings, license validity, or what any AI assistant actually says about a practice.
- Privacy: this page reports aggregate results only; we don't publish per-site results.
Exploratory: solo and group practices
Among checked pages, a credential was found on 55 of 65 solo practices and 45 of 68 group practices, and a named clinician with a role or credential appeared in structured data on 3 of 65 solo and 2 of 67 group practices. Treat this as exploratory: more solo sites couldn't be captured (28 of 101, versus 7 of 80 group sites), so the checked groups aren't directly comparable, and the solo/group labels come with the cohort; we didn't verify them.
Citing the Index
You're welcome to cite these findings with attribution: "Therapist Recognition Index, Second Edition, Strua Practice, September 2026." A link to this page is appreciated. Journalists and researchers can reach Dr. Gehle for methodology questions at [email protected].
Check your own homepage
The same checks behind these numbers are free to run on your own site, without an email. You'll see every check, what it found, and your next steps.
Frequently asked questions
What is the Therapist Recognition Index?
A benchmark of what therapist websites tell people and software, published by Strua Practice. It checks a multi-state sample of private-practice pages with the rule-based Recognition Score and reports aggregate results only. This second edition checked 133 pages captured on September 25, 2026.
What is structured data, and why does it matter here?
Structured data is information a website publishes in its code, usually as JSON-LD, so software can read facts like the practice name, type, location and the people who work there. It isn't part of the visible page. The Index checks whether it's present and what it names; it doesn't test whether any search engine or AI assistant actually uses it.
Can I compare these numbers with the first edition?
No. The first edition used a 0–100 scoring method; this edition uses source checks that report what was found. The two measure different things, so a change between them isn't an improvement or a decline.
Does a missing check mean a therapist is bad at their job?
No. The Index measures what a homepage communicates, not the quality of anyone's clinical work.