Most SEO advice was written for markets where thousands of people search every month. Health practice tech runs on much smaller numbers. A phrase like "practice management software for multi-site dental groups" might pull 90 searches a month in the US, and maybe a dozen of those people run a group big enough to buy anything.
That single fact breaks most of the playbook.
The standard approach assumes attention is scarce and buyers are everywhere. In health practice tech the supply runs the other way. A few thousand groups and practice owners could ever write you a cheque, they mostly know each other, and they search inside a narrow window near the end of a decision that started somewhere else entirely.
So when the traffic chart flattens and your CEO asks what happened, the honest answer is usually that traffic was never the constraint.
Small numbers change what a good month looks like
Run the arithmetic on a low-volume niche and it stops feeling like bad news.
Say one page ranks for a cluster of phrases worth 400 searches a month between them. Sitting at position 2 or 3 gets you somewhere around 60 clicks. If the phrasing is specific enough that only working buyers type it, maybe 15 of those 60 are people with a budget line and a live problem.
Two of those 15 book a demo over the course of the quarter. On a dashboard built for ecommerce that reads as a rounding error. If your average contract is worth $30,000 a year, it's a page that pays for the whole content programme and keeps paying after you stop touching it.
Few teams report it that way. Traffic dashboards reward volume, so the post that pulled 4,000 readers and zero buyers looks like the win of the quarter, and the page that quietly produced two qualified conversations gets cut in the next content audit.
Search happens late in a clinical buying cycle
A practice owner doesn't start on Google. She starts in a WhatsApp group with four other principals, or at a study club, or in a practice managers' forum that has a firm opinion about every vendor in the category.
By the time she types anything into a search bar she already has a shortlist and a suspicion. The search is verification.
Which is why the pages that matter in this market answer verification questions. What implementation actually involves. What happens to existing patient records during migration. Who else her size is running this. What the price looks like once the introductory discount expires.
Keyword research surfaces almost none of that, because verification queries are long, rare, and score terribly in every tool you'd use to justify a content plan.
Four places the standard playbook sends you wrong
1. Chasing the head term
Every tool points at "dental software" or "telehealth platform" because that's where the volume sits. Those results belong to directories, review sites, job boards and the two incumbents who've been publishing since 2012.
The intent is also mixed. A good share of the people searching "dental software" are students and single-chair practices with no budget this year. You can win the term and gain nothing from it.
2. Writing to a word count
Standard advice says match the length of whatever currently ranks. In clinical markets that produces 2,400 words of padding wrapped around 300 words of substance, and a clinical reader spots it inside two paragraphs.
Practice owners read the way people do when they've spent years skimming clinical literature. They scan for the claim, check whether it's supported, and leave. Padding reads as a lack of confidence in the claim.
3. Publishing with no author
Most health tech blogs publish under the company name, or under whoever on the marketing team last had a login. Google's guidance for health-adjacent content asks who wrote a page and why they'd know. Your reader asks the same thing, faster.
A piece about implementation risk written by your head of onboarding, with her name and her actual job on it, does more work than the identical piece published anonymously. Same words, different weight.
4. Clustering around your product
Content plans usually get built around the feature set. One cluster for scheduling, one for billing, one for patient messaging, one for reporting. That's an org chart rendered as a content calendar.
She searches around a problem she's living with this week. "Why patients no-show after the second appointment." "How to hand over a patient list when an associate leaves." Those are the doors into your site, and none of them are named after a module.
Who you're actually competing with in the results
Open the results page for a buying-stage query in this space and count the software companies. Usually one, sometimes two.
The rest are a dental trade publication, a DSO consultant's blog, a Reddit thread, and a 2019 forum post that still ranks because practice owners kept arguing in the comments for two years.
That's the opening. Beating a trade publication on a specific operational question is a one-quarter project for a writer who knows the space. Beating Capterra on "best dental software" is a two-year project with a worse payoff at the end of it.
The competitor set your tool shows you is the wrong one, because the tool is ranking domains by authority score rather than by how well anyone has answered the question.
What to do with the traffic you already have
Most health tech blogs carry a handful of posts pulling decent numbers from the wrong people. A definitions post. A "top 10 tools" roundup written in year one. A patient-facing explainer that a general audience keeps finding.
Deleting them is usually the wrong call, because they hold links and they hold rankings. Repointing them is better.
Take the top five by sessions and ask one question of each: if a buyer landed here by accident, where would you want them to go next? Then put that link in the first third of the page, in a sentence that earns the click, rather than in a related-posts widget at the bottom that almost nobody reaches.
You can do this in a morning. It won't change the traffic number at all, which is the point, and it'll change the share of readers who touch a buying-stage page in the same session.
Do the same audit on your demo request form. If the only route from a useful article to a conversation is a generic contact page, you're asking a cautious clinical buyer to make a bigger jump than she's ready for after one read.
The review cycle that eats your publishing cadence
Something else the generic playbook skips: in regulated markets, publishing has a queue.
Clinical claims go past someone with a licence. Anything touching patient data goes past legal or your privacy lead. If you're selling into the NHS or a US health system, procurement-adjacent claims get a second look too.
Teams plan for 8 posts a month, hit the review queue in week two, and ship 3. Then the cadence breaks, rankings drift, and the channel gets written off as slow.
Two fixes worth the effort. Build a standing claims library, a short document of pre-approved phrasing for the things you say constantly, so 80% of a draft never needs re-reviewing. And book the reviewer's time in advance as a recurring 45-minute slot rather than sending drafts into their inbox and hoping.
Cadence is the part of SEO that compounds. Protect it operationally, or the strategy doesn't matter.
What a version built for this market looks like
Start from the sales call rather than the keyword tool. Pull the last 20 stalled or lost deals and write down the question that killed each one. That list is your first content plan and it costs you an afternoon.
Then check which of those questions people actually type. Some will come back at 40 searches a month. Write them anyway, because the 40 are the right 40.
Two rules keep the output honest.
Every page carries one fact a competitor can't copy. A number out of your own implementation data, a named clinician's view, a detail about how the workflow behaves in a 6-surgery practice on a Monday morning. Something that proves you've been in the room.
Every page carries a real author. Name, role, and one line on why they'd know. If nobody at the company will put their name to a piece, that tells you something about the piece.
Topical authority does the heavy lifting here
In a low-volume niche no single page carries the channel. What carries it is covering one narrow subject so completely that there's no obvious reason to send a searcher anywhere else.
Pick a subject and make it narrower than feels comfortable. "Software migration for growing dental groups" rather than "dental technology." Write the 12 pages that cover it properly, link them to each other in a way that makes sense to a human, and keep them current as the market shifts.
Twelve deep pages on one subject will outperform 60 shallow pages spread across six, and they'll pull better-fit readers the entire time. They also give your sales team something to send mid-cycle, which rarely makes it into the strategy deck and gets used every week.
The number to take upstairs
The reporting problem is that traffic is the easiest number to show and the least useful one you have.
Swap it for four. Sessions from companies that match your target profile. Pages that appear in the touch history of a closed-won deal. Queries where you hold position 1 to 3 among buying-stage terms. And the median gap between a first read and a first sales conversation.
Those numbers move slower, and they hold up against the question a CFO eventually asks about whether any of it produced revenue.
A list of accounts that read four pages before they booked answers that question. Keep the list.