What Actually Goes With Ranking for Therapy Practices: My October 2026 Research

My why behind the research

For starters, besides running HarborVisibility, where I help private practice owners around the world improve their online search visibility to make it easier for the right-fit clients to find them, I'm also training to become a clinical psychologist, which is where I came to see both sides of private practice: the training that prepares you for the clinical work, and the complete absence of training for the part where people have to find you.

Before psychology I trained in computer science at HTL Spengergasse in Vienna, a higher technical college, which is where the technical side of my work comes from. In my work with practices I kept hearing the same advice repeated everywhere, be it "add an FAQ", "write longer pages", "post every week" or "niche down", and at some point I wanted to know which of it actually shows up in the data for therapy practices and which of it just sounds right.

So I built the research myself, and this post is the first report on what it found.

Before you read: what this is and isn't. Everything below is an association, which means it shows which traits go with higher rankings but not what causes them. It isn't a formula and it isn't guru advice. I'm reporting what the data showed, including the parts that didn't hold up when I checked again a week later.

Key findings at a glance

• Across 800 US therapy searches pulled in October 2026, small practices (1 to 5 clinicians) held 8% of top-5 spots, with solo practices holding 5% and practices with 2 to 5 clinicians 3%.
• Directories decide how open a search is. Where practice websites dominated the results, small practices held 14% of top-5 spots, but where directories dominated they held only 2 to 3%. Psychology Today appeared in 79% of the top 10s.
• Distance to the searcher was the strongest factor in every reading. Small practices in the top 5 had a median office distance of 12.7 km from the city's search point, against 18.6 km for those ranked lower.
• Niche focus across the whole site went with ranking, and it held when I re-ran the analysis on completely new sites. Among small practices, a median 26% of a top-5 site's pages were about the searched specialty, against 14% for lower-ranked sites.
• Pages that clearly match the search ranked better, both in what the page text covers and in whether it covers the subtopics the current top 10 cover.
• A title that mirrors the search went with ranking, but modestly, and less strongly a week later.
• Word count, FAQ sections, schema markup, a named clinician on the page, freshness and page speed showed no link to ranking in any reading.
• 72% of the searches small practices win showed zero or no search volume in Google's keyword data (74% a week later), so "no volume" doesn't mean no one searches.
• One ranking check is a weak signal. The same search pulled 21 minutes apart shared only about 5 of its top 10 results.

The full table with a confidence level for each finding is further down.

If you'd like to see how your own site measures up on the things that held up here, you can get a free visibility snapshot.

What I set out to test

The aim: among therapy practice websites competing on the same Google search, which traits separate the ones that rank near the top from the ones further down?

I went in with two beliefs from my own work with therapists:

1. The more you niche down, the better you rank. That's what I'd seen work with the practices I've worked with, but not always. Sometimes a broad practice outranked a focused one, and I wasn't sure which parts of "niching down" were actually doing the work or whether some of it only seemed to matter.
2. The technical foundation is necessary, but it's not what makes you rank. A fast, indexable, mobile-friendly site lets you compete, but I didn't expect it to be what separates the practices that rank from the ones that don't.

How they came out:

• Niche focus was supported, but with more nuance than I expected. What went with ranking was the whole site being about the specialty, starting with the homepage. Simply making the search wording narrower ("EMDR therapist for teens in Austin" instead of "EMDR therapist Austin") barely changed how often small practices got in, by 1 to 2 percentage points, which is within the margin of error.
• On the technical foundation, the second half held and the first half I couldn't test. Sites ranking on page 3 had the basics too and page speed showed no link to ranking, but since almost every site in the sample was technically sound, I can't show that the basics are necessary. I still treat them as the starting point and not the strategy (the order I'd work in is in How to Rank on Google for Therapists).

How I ran the research

Three things are worth knowing about how this works:

• Comparing within a search. A site ranking 3rd for "EMDR therapist Tulsa" is only ever compared with the other sites on that same search, so a search being particularly hard or easy doesn't distort the result.
• How "about the specialty" was measured. I used two ways. The first is simple counting, so whether the specialty appears in a page's address, title or main heading and what share of the site's pages and blog posts that covers. The second is a small open-source AI model (all-MiniLM-L6-v2, running locally on my laptop) that rates how close each page's meaning is to a one-line description of the specialty, which also catches pages that are on topic without using the exact word.
• Locking the rules first. I wrote down what I expected and how I'd test it before pulling any data, and then checked again a week later on new sites with the same locked code. That's the standard way to stop yourself from finding patterns that aren't actually there.

Reading the numbers. Some results are odds, for example "1.5 times the odds of being in the top 5" for a site that's one typical step better on a measure (one standard deviation, roughly the gap between an average site and a clearly above-average one). Where I can, I give plain medians instead because they're easier to picture.

What went with ranking

1. Distance to the searcher was the strongest factor

For local searches, how close the office is to where the search is made was the strongest result in every reading, and about twice the size of any content effect. Small practices in the top 5 had a median distance of 12.7 km, against 18.6 km for those ranked lower, and a week later the sites that kept their top-10 spot were also closer than the ones that dropped out.

What it means for you: if you see clients in person, the searches near your office are your best chances. A practice on the edge of a big city competing for "[specialty] [big city]" is starting from behind, and I'd say that's worth knowing before you judge your website for it.

2. Niche focus across the whole site

Every measure of niche focus pointed the same way, be it the share of pages about the specialty, the share of blog posts about it, the homepage naming it or the AI model's rating of how focused the site is. Among small practices, a median 26% of a top-5 site's pages were about the searched specialty, against 14% for lower-ranked sites, and it held again on the fresh sample of new sites a week later.

There are two nuances though. The measures move together so closely that I can't say which single one matters, so it's really the site as a whole. And across all practice sizes the link was clearest for large group practices, but among small practices on the same search it held for solo practices too.

What it means for you: pick the specialty you most want to be known for and make the site about it, homepage included, since one niche page on an otherwise general site isn't the same thing.

3. The page matches what was searched

Pages whose text was closer in meaning to the search ranked better. In the small-practice study a typical step up went with about 1.95 times the odds of a top-5 spot, and 1.58 times on the fresh sample, and in the re-check it was also supported across all practice sizes.

What it means for you: the page you want to rank should say plainly what it's about, who it's for, where you work and what someone should do next.

4. Covering what the current top 10 cover

I grouped what the top 10 pages for each search talk about into subtopics, and small-practice pages that covered more of those shared subtopics were more likely to be in the top 5 (1.51 times the odds per step, 1.31 on the fresh sample). Page length itself showed no link at all.

What it means for you: look at what the pages ranking for your search actually answer and make sure yours answers those things once and clearly. More words don't help, but covering what a reader expects to find does.

5. A title that mirrors the search, modestly

Among small practices, a page title closer to the search wording went with ranking in all three readings (1.69, 1.60 and 1.35 times the odds). Across all practice sizes it was significant the first week and just short of it in the re-check, and for solo practices specifically it pointed the right way in 5 of 6 readings but was only significant in 2.

What it means for you: title your page with the words people actually search. It's cheap and consistent, but it's a small signal, so I wouldn't expect it to move you on its own.

6. Directories decide how open a search is

Small practices held 8% of all top-5 spots, and that share swung a lot with the type of results page: 14% where practice websites dominated, 8% on mixed pages, and only 2 to 3% where directories and institutions held 6 or more of the top 10. Psychology Today appeared in 79% of the top 10s, Reddit in 28% and Facebook in 23%.

What it means for you: search your main term and count the directories in the top 10. If there are 6 or more, your directory profile is doing more of the work for that search than your website can. That's not a reason to give up on the website, it's a reason not to rely on just one source.

7. AI Overviews cite what already ranks

Google showed an AI Overview on 58 of the 168 main-study searches, and practice sites ranked 1 to 5 were cited in it 46% of the time, while sites ranked 21 to 50 were cited only 4% of the time (52% and 5% a week later).

What it means for you: for Google's AI answers, ranking well is the way in. I didn't see any separate trick in this data.

What showed no link to ranking

These are often sold as ranking work, but in my data none of them went with better ranking in any reading once domain age and backlinks were held constant:

• word count and number of subheadings
• FAQ sections and "what to expect / how it works / who it's for / fees" checklists
• schema markup (structured data)
• a named, credentialed clinician on the page
• an address or map on the page
• how recently the page was updated
• how many blog posts a site publishes
• page speed scores

That said, many of these are still worth doing for the person actually reading your site, since a clear fees section or a real bio helps someone decide to contact you. They just didn't separate the sites that rank from the ones that don't.

What didn't hold up a week later

I re-ran everything on October 12 with the same locked analysis, and four earlier results didn't hold:

• "Pages linked more prominently inside a site rank worse." This surprising first result disappeared.
• "Wix sites rank worse." It weakened to a lean that's no longer significant.
• "Niche focus matters twice as much for online-only practices." This wasn't repeated.
• "Homepages never win generic searches like 'therapist [city]'." A week later, 3 of 4 generic winners were homepages.

I'm including these on purpose, because if a result can flip within a week, it was never solid enough to build advice on in the first place.

One more thing the re-check showed is how much a single ranking check moves. The same search pulled 21 minutes apart shared only about 5 of its top 10 results, about 3 hours apart it was 4.6 and a week apart about 4. Whether a results page counted as "directory-dominated" even flipped between same-hour pulls in 5 of the 12 markets I tested. So one search on your phone tells you very little about where you actually rank, and Search Console's average over many real searches is a much better guide.

Solo vs group practices: what I know so far

My reading so far: for a solo practice, being close to the searcher and having a site that's about one specialty, homepage included, are the best-supported things. Whether being a group is itself an advantage is still the open question, and it's the main thing the next round tests. For what group practices need to set up differently in the meantime (Business Profiles, provider pages), see Does Group Practice SEO Work Differently Than Solo Practitioner SEO?

Summary table

Limits

• Association, not cause. Something I didn't measure could drive both a trait and ranking, for example reviews, how often people search a practice by name, how they click or how strong a practice's Google Business Profile is.
• Scope. These are US searches in English on desktop, for a handful of therapy specialties and in one month (October 2026). Google's September 2026 spam update was also still rolling out during the first pulls and only finished on October 8.
• Some measures are rough. "About the specialty" partly relies on the term appearing in addresses, titles and headings, and the AI model is small and general-purpose.
• Not independently reviewed. I designed and ran this myself, with Claude's help on the code and analysis. The locked code makes it reproducible, but no one outside has checked it yet.

What's next: this research is ongoing

I'll keep this page updated as new results come in, and there are two things running right now:

• A weekly ranking experiment (from October 28, for 8 weeks), where 40 new searches are each pulled 3 times every week. The question is whether a one-week move in ranking predicts where a site stands four weeks later or whether moves mostly reverse, and I expect results in mid-December.
• Solo vs group, November and December 2026, with about 300 new searches across 122 US cities, each pulled 3 times. It tests whether group practices outrank solo practices on the same search and whether more clinicians goes with ranking, and December then repeats the test on a new set of searches so that a single result can become a confirmed one.

Update log

• October 29, 2026: Version 1, October 2026 data and the October 12 re-check.

How to cite this research

Otter, M. (2026). What actually goes with ranking for therapy practices: October 2026 research. HarborVisibility. https://harborvisibility.com/insights/therapy-practice-ranking-research-2026

Frequently asked questions about this research