Choosing Medical Practice Marketing Services: A Guide

Ratings move stated preference. In a controlled experimental survey of primary-care physician selection, moving a rating from two stars to four stars increased the relative log odds of a patient choosing that physician by 1.31 for government clinical ratings and 1.32 for commercial nonclinical ratings ¹. The two source types pulled almost identically, which tells a practice owner something specific about what patients respond to: participants weighted a familiar commercial rating as heavily as an official clinical score.

The scope matters. The study measured hypothetical choice in a survey environment, not completed appointments and not clinical outcomes ¹. A four-star profile does not prove better care. It predicts that when a prospective patient lands on a comparison view with the practice next to two competitors, the higher-rated profile wins more clicks. That is a demand-generation lever, not a quality credential.

An agency proposal that treats rating lift as the deliverable is selling the top of the funnel and calling it the whole funnel. A defensible reputation program uses the rating effect for what the evidence supports: increasing the probability that a searcher selects the practice’s listing over an adjacent one. It should be paired with review volume growth and with tracking that follows those clicks into calls and scheduled intakes, so the agency can show what the star lift actually produced downstream rather than reporting the star number as the outcome itself.

The credibility question has been studied directly. A 2023 systematic literature review of physician-rating websites included 28 studies after screening 36,755 candidates. Seven supported the credibility of ratings against alternative measures, six found no correlation, and 15 reported mixed results ³. That distribution is the practical answer to any agency claim that star averages substitute for a quality report.

The broader evidence base moves in the same direction. A systematic review of 32 studies found consistent positive correlations between online reviews and patient experience, and mixed associations with clinical quality measures, with organizational-level reviews generally more reliable than reviews of individual clinicians ². What patients write about most often reflects communication, wait time, staff interaction, and perceived responsiveness rather than validated clinical endpoints ⁴.

For a mental health practice, that has direct reporting consequences. A reputation dashboard should treat reviews as an experience instrument: sentiment themes, response time to negative reviews, share of reviews mentioning intake friction or scheduling, and the volume trend that separately drives visits ⁵. It should not be framed as evidence that the practice delivers better care than a competitor with a lower average. An agency that conflates the two is either underinformed about the literature or comfortable letting a client repeat a claim the evidence does not support.

Impressions, sessions, and average position describe traffic. They do not answer whether the practice added intakes this quarter. A measurement stack worth paying for tracks each inquiry from source to scheduled appointment, and reports the numbers a clinical director can defend to partners.

Four counts do most of the work:

1. Qualified inquiries by source: calls, form fills, and secure-message requests that meet a written definition (in-service, in-market, insurance or self-pay match, adult or pediatric as applicable).
2. Inquiry-to-intake rate by source, which exposes whether a channel is delivering people who actually complete the intake process.
3. Intake-to-first-appointment rate, which catches drop-off between scheduling and the first session.
4. Review volume trend, tracked separately from average star rating because volume growth has been shown to move patient decisions and outpatient visits independently of the score itself ⁵.

Two secondary measures give the numbers meaning: cost per qualified inquiry by source, and completed appointments per source over a defined lookback. Together they let a practice reallocate budget away from channels that fill the top of the funnel without producing scheduled care.

Two guardrails belong on the same dashboard. Any conversion tracking on pages that concern specific conditions, providers, or appointment requests must be reviewed against HHS Office for Civil Rights guidance on tracking technologies before deployment, including vendor business associate coverage ⁸. And the definition of a qualified inquiry should be written down and version-controlled, so a quarter-over-quarter comparison is not a moving target.

Health content on the open web sets a low bar. A meta-analysis of 11,785 health information websites found that only 18% carried Health On the Net Code certification, and using the DISCERN instrument, no site was rated excellent while between 37% and 79% were rated good depending on the topic and specialty ⁷. The distribution establishes the baseline a mental health practice is competing against and inheriting from when an agency drafts pages about anxiety, depression, medication management, or intensive outpatient options.

Three production controls close the gap:

1. A named clinical reviewer with license number and specialty attached to each substantive page, not a generic “medically reviewed” badge.
2. An author byline that identifies whether the writer is a staff clinician, a credentialed medical writer, or a contract copywriter, so a reader and a regulator can assess who authored a health claim.
3. An editorial workflow that tracks each factual assertion to a primary source, with citations visible to the reader rather than hidden in a style guide.

An agency proposal worth signing describes the reviewer bench, the turnaround for clinical sign-off, the citation standard, and the version control that shows what changed after review. If the sales conversation cannot answer those four questions, the practice is buying volume, not accuracy, and the DISCERN and HON distributions predict where its content will land.

Language choices in mental health copy shape whether a prospective patient calls or closes the tab. Patient-authored reviews of clinicians cluster around communication, staff interaction, and perceived responsiveness rather than clinical endpoints ⁴, and page copy that mirrors those concerns tends to move the same readers who write those reviews. Copy that reads as diagnostic, judgmental, or clinical-textbook cold does not.

Three writing standards belong in the agency’s style guide before any page is drafted:

• Person-first construction as the default (“a person living with bipolar disorder,” not “a bipolar patient”), with identity-first phrasing used only where a specific community prefers it.
• Plain-language reading level tied to a measurable target, since accessibility is one of the DISCERN quality dimensions applied to health information ⁷.
• Explicit prohibitions on outcome language the practice cannot substantiate, including recovery-rate claims, timelines to symptom relief, and comparative superiority claims.

A practice owner reviewing sample work should read a full landing page aloud. If it sounds like an insurance brochure or a diagnostic manual, it will underperform copy written by someone who has sat across from a person in the waiting room. The agency’s edit log should show clinician comments that changed word choices, not just fact checks.

Owned content carries a promotional discount in the reader’s mind. Research on which sources consumers trust when evaluating physician quality identifies three source clusters that carry weight beyond the practice itself: employers and healthcare providers, user-advocacy organizations, and insurance companies and government agencies ⁶. Content that cites into those clusters imports credibility a practice page cannot manufacture on its own.

The practical standard is a mixed citation pattern. A page describing evidence-based treatment for major depressive disorder should link to guidance from federal health agencies or professional bodies. A page on insurance coverage should cite the specific payer’s medical policy or the state parity statute. A page on telehealth access should cite HHS resources on cross-state practice rules ^{11, 12}. Practice-generated claims sit alongside those anchors, not in place of them.

An agency that resists outbound citations because they “send traffic away” is optimizing for a metric the practice does not need. Trust conversion, not session duration, is what fills the intake calendar, and the source clusters patients already rely on are the shortest route to it.

The tracking layer is where most agencies quietly transfer risk to the practice. HHS Office for Civil Rights guidance is explicit that regulated entities may not use tracking technologies in a way that causes impermissible disclosure of protected health information to tracking vendors, and that when a vendor creates, receives, maintains, or transmits PHI for a covered function, that vendor may be a business associate requiring a business associate agreement ⁸.

A defensible diligence pass touches six surfaces:

• Analytics and advertising pixels on pages that concern specific conditions, providers, or appointment requests.
• Session replay tools that can capture form entries or free-text symptom fields.
• Appointment and intake forms, including any pre-screen that collects reason for visit.
• Patient portal login pages and post-login analytics.
• Chat widgets and their transcript storage.
• The vendor contracts behind each of these, including which vendors will sign a BAA and which will not.

Two nuances belong in the conversation. A federal district court vacated the bulletin’s application to one category of unauthenticated public-page visits involving an IP address connected to a page addressing specific health conditions or healthcare providers, and HHS stated it was evaluating next steps; the remainder of the guidance remains in force ⁸. Cookie banners, privacy-policy language, and downstream de-identification do not automatically resolve the underlying disclosure question.

HIPAA defines marketing as a communication about a product or service that encourages recipients to purchase or use it, and with limited exceptions, covered entities generally need written authorization before using or disclosing PHI for marketing ⁹. That definition reaches further into a mental health practice’s campaign stack than most agencies acknowledge.

Three activities warrant a direct authorization conversation before launch:

• Uploading a patient email list to an ad platform to build a remarketing or lookalike audience, which discloses the fact of the treatment relationship to the platform.
• Running email or SMS campaigns to prior patients that promote add-on services, group programs, or ancillary offerings.
• Publishing patient testimonials, video stories, or before-and-after narratives that identify the individual as having received care at the practice.

Written, revocable authorizations with a specific description of the disclosure, the recipients, and an expiration date are the operational answer. The agency should be able to name which of its proposed tactics require authorization, which fall inside a treatment or health-care-operations carve-out, and which get routed through the practice’s compliance officer before the first send.

Outcome copy is where compliant marketing quietly turns into an enforcement problem. The FTC standard is that advertising must be truthful and not misleading, and that advertisers must have adequate substantiation for objective claims before dissemination, with health-related benefit and safety claims generally requiring competent and reliable scientific evidence ¹⁰. That standard applies to the totality of the ad’s impression, not just the words in the headline.

Practical language for a mental health practice: recovery-rate percentages, timelines to symptom relief (“feel better in two weeks”), comparative superiority (“the most effective program in the region”), and success-story testimonials that imply typical results all trigger substantiation and disclosure obligations ¹⁰. Implied claims from imagery, testimonial selection, or star-metric juxtapositions count the same as express claims.

Before signing, ask the agency to show its claim-review workflow: who drafts, who substantiates, what counts as adequate evidence for each claim type, and where disclosures are placed. If the process is informal, the practice is one campaign away from a claim it cannot defend.

Geographic targeting is a licensure statement whether the agency treats it as one or not. HHS guidance for behavioral health is direct: providers must be licensed or legally permitted to practice where the patient is located, and before providing telebehavioral health, providers should verify the licensing boards where both the provider and the patient are, professional-liability coverage, and reimbursement requirements ¹¹.

Cross-state work is not a single question. Possible pathways include:

• Obtaining a full license
• Checking temporary-practice laws
• Using reciprocity
• Joining a licensure compact such as PSYPACT or the Counseling Compact
• Applying for telehealth registration where available ¹²

Each pathway carries different scope, duration, and prescribing rules, and states continue to revise them.