SEO for doctors: the highest bar in search, and what clears it

Google's rater manual sorts health questions by how much damage a wrong answer can do. Among its worked examples, "when to go to the emergency room" is clear YMYL, the category whose pages "require the most scrutiny for Page Quality rating"; "how often to replace a toothbrush" only may be YMYL, casual enough to ask a friend. Medicine is the example Google reaches for whenever it needs to show what its strictest standard looks like, and that reverses the usual framing: for most local businesses search is a visibility problem, while for a medical practice it is a credibility problem first, and the credential the practice already holds is the asset almost every guide leaves unused.

Quick answer

• Health is YMYL territory. Google's rater guidelines hold medical pages to their highest published quality standard.
• The verifiable credential decides, not the volume. An unsigned health article is the guidelines' textbook Low rating.
• The condition is the unit of strategy. Most patient search paths start with a symptom or a condition, not with a doctor's name.
• Reviews carry a constraint no other sector has. A reply that confirms anything about a patient is a privacy violation with an enforcement record.

What does SEO for doctors involve?

SEO for doctors is the work of being findable for the conditions and treatments a practice handles, under the strictest quality standard Google applies to any topic. A solo practitioner clears it by signing every page; a multi-specialty clinic by naming a reviewer per specialty; a hospital group by consolidating the duplicate pages its departments already published.

The mental model that fails here

The plan most practices buy is retail logic with medical words. A blog on a weekly schedule, an article per keyword, volume as the road to authority. Under a YMYL standard that plan does not merely underperform: anonymous volume is close to the specific thing the standard exists to filter out. The rater guidelines rate a flu-duration article Low because there is "no evidence that the author has medical expertise", and their Lowest medical example is a site with "literally no information about who is responsible for the content". In both cases, the failure named first is not the medicine. It is that nobody checkable stands behind it.

The sector next door sharpens the contrast. A dental practice is decided mostly by insurance networks, capacity and proximity, a different problem with a different plan, covered in SEO for dentists. A physician's content competes upward instead: on almost any condition query the other results are hospital systems and national reference sites, and the one lever a small practice holds against them is the credential those competitors cannot localize. At group scale the problem changes shape again, because the entity that ranks and the entity that decides stop being the same person: SEO for healthcare.

What Google documents about medical content

The documentation is unusually explicit. Two public documents set the frame.

The Search Quality Rater Guidelines, 182 pages in the version dated September 11, 2025, define YMYL topics as those that "could significantly impact the health, financial stability, or safety of people, or the welfare or well-being of society", and state that "for pages about clear YMYL topics, we have very high Page Quality rating standards". Their example of a query where "mild inaccuracies" already matter is medical: symptoms of a heart attack. The introduction settles the tone in one line: "Medical search results should be high quality, authoritative, and trustworthy."

The guide to creating helpful content connects that manual to the ranking systems: "our systems give even more weight to content that aligns with strong E-E-A-T for topics that could significantly impact the health, financial stability, or safety of people". Money sits inside that sentence alongside medicine, which is why the same named-author requirement governs accounting content, where the constraint plays out around a tax calendar instead of a clinic schedule and is covered in SEO for accountants.

One precision, because the current first page of results tends to skip it: rater ratings are not rankings. The same guide states that "rater data is not used directly in our ranking algorithms". The guidelines matter as the most detailed public statement of what those systems are tuned to reward, and for medicine they are specific about the failure they punish: content whose author cannot be identified or whose expertise cannot be checked.

What experience may say, and what is reserved for expertise

The guidelines draw a line inside health content that almost no medical marketing plan draws. Their own paired example: a forum where people describe coping with liver cancer treatment is valuable personal experience, while "different treatment options for liver cancer and the associated life expectancies under each treatment" is information best left to experts. A practice can publish both kinds of page; deciding which is which decides who signs what, and it separates a patient-stories section that builds trust from one that quietly commits the practice to clinical claims nobody reviewed.

The condition is the unit, not the biography

Patients search around problems. Practices build websites around people, and the mismatch is measurable.

The oldest study this vertical still quotes is clearest exactly here. Google and Compete surveyed 533 hospital researchers in May and June 2012 and classified their search paths: 37 % started with symptoms or departments, 38 % with conditions or diseases, 6 % with treatments or procedures, and 19 % with a brand. The study's own summary is that most paths begin with a non-branded term. Those percentages are fourteen years old and belong to hospital selection, so they travel with their date; the shape they describe is still what shows up in a Search Console fed by a medical site.

One page per condition or treatment

A services page that lists everything the practice treats competes for none of it. For a solo practitioner the unit produces a short list: the handful of conditions that actually fill the schedule, each with its own page. A multi-specialty clinic multiplies that by specialty and needs an owner per set, because unowned pages are the ones that go stale. A hospital group usually faces the reverse work: departments have published near-duplicates of the same condition page for years, and consolidating them moves more than writing anything new.

The winnable query is specific. Not the naked condition, where the ceiling belongs to national reference sites, but the condition attached to an intent those sites cannot serve: availability, locality, what a first appointment involves, what a procedure costs at this practice, which insurance applies. Those are questions a clinic answers with an authority no encyclopedia has.

What the biography page is actually for

It ranks for the physician's name, and that is a real job rather than the strategy. A referred patient searches the name before calling, and the page confirming the license, the board certification and the affiliations converts a decision already half-made. Its second job is quieter: it is the proof that the author and reviewer lines across the condition pages point to a real, checkable person.

The reviewer line is the credential made visible

The pattern that clears the bar costs one line per page: written or medically reviewed by a named clinician, credential stated. In the United States that claim is checkable in public records, through the state medical board registries and the NPI registry, which turns a byline from a marketing claim into a verifiable fact. The property cannot be manufactured, and the rater guidelines reserve some of their harshest language for "a creator blatantly misrepresenting their medical credentials". The real constraint is clinician time: a review hour is an unbilled hour, which is why the honest version of this strategy publishes fewer pages than the marketing plan wants. Each reviewed page is one a competitor without clinicians cannot copy.

Reviews, asked for and answered under HIPAA

In healthcare the reply is regulated. That single fact reorganizes the entire review playbook.

HIPAA treats a public reply as a disclosure, and the enforcement record is concrete. In June 2023 the US Department of Health and Human Services settled with Manasa Health Center, a New Jersey psychiatric practice, for 30,000 USD after replies to negative Google reviews disclosed the diagnoses and treatment details of four patients. The earliest settlement of the kind came in 2019, when Elite Dental Associates paid 10,000 USD over a reply to a patient's public review.

The operational consequence runs deeper than tone. A compliant reply cannot confirm the reviewer was ever a patient, cannot mention a visit or a condition, and cannot correct the facts, even when the review is false and the reviewer told the story first. The rebuttal every practice wants to write is the one thing it cannot publish. What remains is arithmetic: a steady flow of recent reviews that outweighs any single bad one, which moves the effort from answering reviews to generating them. The general flow is in how to get Google reviews; the healthcare adjustment lives in the asking. A request sent uniformly to every patient, with no mention of why they were seen, stays on the safe side. Selecting recipients by outcome or by procedure is a use of clinical information the marketing tool was never authorized to make.

The local layer still decides the visit

The visit is still local. The YMYL bar governs what the pages may claim; proximity and the profile govern who walks in.

For the "specialty near me" family of queries the machinery is the same as for any local business, and it starts with the profile, worked field by field in optimizing your Google Business Profile. One rule is specific to this sector and worth knowing: Google's guidelines let an individual practitioner in a public-facing role hold a listing of their own, under their name alone, separate from the practice's. A group whose physicians go unlisted competes with one profile against clinics that field one per doctor plus their own.

The map itself is ordered by the three factors Google documents, relevance, distance and prominence, and the mechanics of moving the one that moves are in how to rank in Google Maps. Distance is fixed by the address. Prominence leans on reviews, which the previous section just constrained. What is left is relevance, which is written, and that is the same conclusion the condition pages already forced.

Everything else in the local layer, categories, hours, photos, consistency across directories, follows the general playbook in local SEO for small business, with the healthcare constraints layered on top rather than replacing it.

When not to invest in this

When no clinician has review time. A health article published without clinical review is not a head start on the strategy; it is the exact artifact the quality systems are tuned against, and retiring it later costs more than never publishing it.

When the schedule already fills from referrals and the panel is closed. Ranking produces demand somebody has to absorb, and what this channel can and cannot do for a business that does not need demand is the subject of what SEO is and what it is for.

When the money is needed this quarter. The clock of this channel does not accelerate for medicine, and the competition raises the price: what the market bands realistically buy is in how much SEO costs for a small business.

When the practice's real problem is capacity, phones or intake. More visibility multiplies an operational failure. It never fixes one.

The number every guide quotes is fourteen years old

The figure is real, and it is old. The first page of results for this query repeats that 77 % of patients used search before booking an appointment, and none of the pages that quote it say when it was measured. The source is the 2012 Google/Compete hospital study, a survey of 533 hospital researchers run in May and June 2012, with the year printed on its cover.

Age does not make the number false. It makes it a historical measurement offered as a current one, in a market that has since gained AI summaries, zero-click results and telehealth. A practice will be shown that figure, or one of its cousins, in a sales meeting this year, and the useful response is not general doubt but two questions with checkable answers: measured when, and over whom. A vendor's reaction to those two questions says more than the deck does, and what else belongs in that meeting is in choosing an SEO agency. The finding actually worth stealing from that study, that most paths began with a non-branded term, almost never gets quoted: it argues for architecture work instead of for a budget.

Mistakes that repeat

• Publishing health content nobody signs. The rater guidelines' Lowest medical example is a page where no one is responsible for the content.
• Replying to a review with the facts of the visit. The reply itself is the violation, and practices have paid five-figure settlements for it.
• Building the site around the biographies. Most patient search paths start with a symptom or a condition, and a biography answers neither.
• One services page for every condition. It competes for nothing, in a vertical where the competition is hospital systems.
• Quoting the 77 % figure as current. It was measured in 2012; a guide that offers it undated is showing you its own sourcing standard.
• Hiding the credential the practice already has. The license is the one signal a content operation cannot fake, and it is useless while it stays off the page.

Data and transparency

The YMYL definition, the "very high Page Quality rating standards" for clear YMYL pages, the Low and Lowest medical examples and the experience-versus-expertise distinction come from Google's Search Quality Rater Guidelines, version dated September 11, 2025, opened on 18 August 2026. That systems "give even more weight" to strong E-E-A-T on health topics, and that rater data is not used directly in ranking, comes from Google's guide to creating helpful content, opened the same day. That an individual practitioner may hold a separate profile comes from Google's guidelines for representing your business, also opened that day.

The 77 % figure and the search-path split come from the 2012 Google/Compete hospital study, a survey of 533 hospital researchers run in May and June 2012; both are cited with their date because the age is part of the argument. The Manasa Health Center settlement (30,000 USD, four patients, June 2023) and the 2019 Elite Dental Associates settlement (10,000 USD) sit on the public enforcement record of the US Department of Health and Human Services' Office for Civil Rights; the HHS announcement pages refuse automated access, so the figures were checked against HIPAA Journal's report, opened on 18 August 2026. The pricing reference in the FAQ comes from SE Ranking's survey published 13 December 2024, 260 agencies, whose authors warn the sample may not reach statistical significance.

No figure appears here for cost per click in this vertical, for patient lifetime value, or for the share of patients who read reviews: versions of all three circulate widely and none I checked traced to a primary source. The strategic ordering is operating judgment from audit work across a portfolio recording more than 300 million impressions a year in Search Console. Nothing here is legal advice, and the specifics of any practice's privacy obligations are for the practice and its counsel to confirm. Verified as of August 2026.

Primary sources, opened on 18 August 2026: the Search Quality Rater Guidelines (PDF); Google's guide to creating helpful content; Google's guidelines for representing your business; the 2012 Google/Compete hospital study (PDF); HIPAA Journal's report on the Manasa settlement; SE Ranking's pricing survey.

What this changes

The bar is the moat. Everyone arrives at this query asking how to survive the scrutiny Google applies to health content, as though the standard were a tax on the sector.

Run the logic the other way. The standard filters for verifiable medical credentials, and a licensed practice is one of the few participants that holds them by default. A content operation can outproduce any clinic forever; it cannot produce a license number that resolves in a state registry. Every notch the bar rises removes competitors who write more and can prove less. The practices that lose under YMYL are overwhelmingly the ones that hid their strongest asset behind an unsigned blog, which means the work is not manufacturing authority. It is making the authority the practice already has checkable, page by page, name by name.

Frequently asked questions