Behavioral Health Marketing: From Growth to Proof

The category crossed a line most operators are still absorbing. Among commercially insured individuals, behavioral health visits totaled 66.4 million in 2024, compared with 62.8 million primary care visits ⁷. That is not a rounding difference. It is a reordering of which service line commands the largest share of ambulatory demand in a payer segment that has historically defined “mainstream” utilization.

The composition of that demand is equally instructive. Behavioral health accounted for 67% of all telehealth encounters in 2024, meaning two of every three virtual visits in the commercially insured population were mental health or substance use encounters ⁷. Primary care, urgent care, and specialty care collectively made up the remaining third. Virtual delivery is now, in practical terms, a behavioral health infrastructure.

For growth leaders, the operational reading is direct. Search demand, referral flow, and payer contracting no longer sit downstream of primary care marketing budgets. A prospective patient with anxiety, an alcohol use disorder, or a teen in crisis is not routing through a family physician first at the rates that dominated pre-2020 planning. Campaigns that still treat behavioral health as a specialty niche are pricing media and staffing intake against an outdated volume assumption.

Category leadership requires category-appropriate spend allocation, intake capacity, and geographic coverage modeling calibrated to 66.4 million visits, not to legacy specialty benchmarks.

Telehealth is not the growth line it was two years ago. Virtual behavioral health visits among commercially insured patients dropped from roughly 50 million in 2023 to 40 million in 2024, a 20% year-over-year contraction inside a category that is otherwise growing ⁷. The tapering is not confined to commercial books of business. CMS data on Medicare Part B beneficiaries shows telehealth utilization declining from 48% in 2020 to 34% in 2021, 29% in 2022, and 24% in 2023 ⁵. Two independent payer populations, one story: the pandemic-era ceiling has come down, and it is settling well below peak.

Two variables explain most of the movement. Clinical preference has reasserted itself for higher-acuity intake, medication management with controlled substances, and family-involved treatment. Payer policy has also tightened, with cross-state licensure friction and shifting parity rules narrowing where virtual sessions actually get reimbursed.

For media planning, this closes the door on telehealth-only positioning. A campaign that routes every intent signal to a virtual intake form leaves in-person demand unmet at the exact moment when the in-person share is recovering. Landing pages, Google Business Profile listings, and paid search extensions must present hybrid options at the same level of prominence, with location-specific proof points rather than a national telehealth pitch.

Operators running multi-state footprints should audit their intake routing this quarter. If a caller with a preferred in-person modality lands in a telehealth queue by default, the leak is measurable in booked-appointment rates, not just campaign CTR. CMS has separately encouraged insurers to notify beneficiaries of telehealth availability, which means access messaging still matters ¹⁷, but the marketing thesis has shifted from “virtual replaces in-person” to “virtual complements a defended physical footprint.”

Access proof answers a narrow question: did a person who needed care actually reach a clinician? Not clicks, not form fills, not chatbot sessions. A booked assessment with a licensed provider, in the modality and location the patient chose. That distinction matters because public attitudes still favor professional support over unguided digital tools, and self-help is not the preferred choice for handling mental health problems for most individuals ³. Campaigns that route intent toward standalone apps or self-guided programs are marketing against demonstrated preference.

SAMHSA frames digital therapeutics as tools that can extend the reach of evidence-based treatment into preferred environments ¹⁶, which is the correct positioning: digital as an on-ramp to a clinician, not a substitute for one. CMS made the same operational point when it encouraged insurers and health plans to promote telehealth and notify beneficiaries of its availability ¹⁷. Access marketing has a payer-adjacent role, not just a direct-to-consumer one.

The KPI that operationalizes access proof is intake-to-assessment conversion, measured by geography and modality. A landing page that generates 400 form submissions in a rural county but produces 12 completed assessments has an access problem, not a traffic problem. VPs should segment the funnel by ZIP code, insurance status, and requested modality, then compare completed assessments per 1,000 population against the segments the organization actually serves. Geo-level intake call rates and appointment-hold-through-first-session rates are the operator-defined variables that turn access from a slogan into a number the finance team recognizes.

Trust in online mental health services is multi-factorial. A 2024 study on trust in online psychotherapy identified four consistent drivers: the specialist, the platform, education, and the payment process ¹⁹. None of them are creative concepts. All four are operational assets that either exist on a provider’s site or don’t.

The specialist factor rewards clinician-level bios with credentials, licensure states, and photographs of real people. Rosters that read as anonymous “our team of experts” copy underperform against pages that name the licensed clinical director, list the LCSWs and LMFTs by specialty, and disclose training. Platform trust is a proxy for technical quality: page load performance, mobile responsiveness, form security indicators, and clear privacy language. Education means substantive content about conditions and treatment approaches, not blog posts optimized purely for search intent. Payment process trust is the least discussed and most fixable: transparent insurance verification, upfront statements about self-pay ranges when applicable, and a scheduling flow that does not require a credit card before a human conversation.

Trust proof is measurable. Branded search share, direct-traffic ratio, review response latency, and the percentage of intake calls originating from pages with clinician bios versus generic service pages are all operator-defined KPIs a VP can report quarterly. When a marketing team can show that pages with named clinicians convert at a higher rate than anonymized equivalents, the trust argument stops being philosophical and becomes a media-mix decision.

Continuity is where most operators leave admissions on the table. Among Medicaid and CHIP beneficiaries treated for a substance use disorder or mental health condition in an inpatient or residential setting during 2023, 51% received at least one service in an outpatient or home or community-based setting within 30 days of discharge, and 39% received two or more services in that window ⁶. Roughly half of discharged patients do not appear in outpatient follow-up at all. That gap is not a clinical mystery; it is a marketing and operations failure the industry has normalized.

For multi-site behavioral health portfolios, the continuity gap is also a growth lever hiding in plain sight. Every discharged patient without a scheduled step-down appointment represents a lost readmission opportunity into the organization’s own outpatient, IOP, or medication management line. The economics favor internal continuity: a patient already assessed, insured, and clinically engaged does not need to be reacquired through paid search.

Continuity proof is measurable and attributable. The KPI is 30-day post-discharge engagement rate, benchmarked against the 51% Medicaid figure and segmented by discharging facility, level of care, and payer. Discharge planners, alumni outreach, and marketing-automation flows that deliver appointment reminders, transportation resources, and family-support content within the first 72 hours after discharge move the number. VPs who own admissions volume but not census retention are measuring the wrong denominator. Continuity marketing feeds inpatient volume the following quarter through referral loops that outpatient graduates create when they stay engaged.

The specifics matter. The March 2023 FTC order against BetterHelp did not issue a general warning about privacy. It banned the company from sharing consumers’ personal information with certain third parties for retargeting—the practice of serving ads to people who had previously visited BetterHelp’s website or used its app ⁹. The banned data was mental health engagement data: email addresses, IP addresses, and questionnaire responses passed to advertising platforms and used to build custom and lookalike audiences.

Read that scope carefully before assuming a workaround exists. Any behavioral health operator running a Meta pixel, a TikTok pixel, or a Google Ads remarketing tag on pages tied to a mental health intake questionnaire, a self-assessment quiz, an SUD screener, or a treatment-inquiry form is running the exact configuration the FTC found problematic. The consent BetterHelp obtained through its intake flow was deemed insufficient because it did not clearly disclose the advertising use.

Most operators conflate these two rule sets and pay for it during audits. HIPAA and 42 CFR Part 2 apply to different data, require different consents, and carry different exceptions. Marketing teams handling both mental health and substance use disorder programs need to treat them as parallel frameworks, not a single compliance workflow.

Under HIPAA, marketing is defined as a communication about a product or service that encourages purchase or use, and it generally requires prior authorization from the individual ¹². The rule carves out exceptions for treatment communications and certain care-coordination messages, which is why a discharge planner can call a patient to schedule a follow-up appointment without a separate marketing authorization ¹³. Covered entities and business associates must still obtain valid HIPAA authorization to use or disclose PHI for marketing outside those exceptions, and that obligation now overlaps with FTC enforcement on consumer health data collected through websites, forms, and pixels ¹⁴.

42 CFR Part 2 sits on top of HIPAA for substance use disorder records. It is stricter. A written consent to disclose SUD treatment information—including for fundraising or promotional communications—may be paper or electronic, but it must include specific terms governing the uses or disclosures and the patient’s right to elect not to receive such communications ¹⁰. Persons subject to the regulation must comply with the applicable requirements of the final rule by February 16, 2026 ¹¹.

The operational split is concrete. A general mental health outpatient program can, with a valid HIPAA authorization, use a patient’s information in marketing workflows that fall outside the treatment exception. An SUD program cannot fold that same patient into an alumni email campaign, a testimonial video, or a referral outreach list without a Part 2–compliant consent that specifies the exact use. Admissions CRMs that mix mental health and SUD records into a single audience segment are the highest-risk configuration in most portfolios.

The FTC’s health products compliance guidance treats brochures, website copy, social posts, influencer content, and press releases as advertising, and it applies the same substantiation standard across all of them ¹⁵. Outcome claims about behavioral health treatment—success rates, sobriety percentages, symptom reduction figures, average length of stay to remission—need competent and reliable evidence tied to the population and conditions the claim describes.

Testimonials carry the same weight. A patient’s statement that a program worked for them is an advertising claim about typical results unless the material clearly discloses otherwise, and the underlying claim still requires substantiation. Before-and-after language, recovery-rate infographics, and “90-day success” statistics without a cited internal outcomes study or a published source should be rewritten or removed. Marketing teams should keep a substantiation file for every quantitative claim currently in market, dated and tied to a data source the clinical or quality team owns.

A pilot study of youth entering mental health treatment for mood and anxiety disorders found that most who self-diagnosed pointed to social media as a contributing factor in that belief ¹⁸. That data point should sit on the desk of every content lead approving symptom-list articles, TikTok scripts, and quiz-style landing pages. Copy that reduces complex diagnoses to three checkbox symptoms produces qualified-looking traffic and unqualified clinical encounters.

The evidence on digital-first engagement adds a second caution. In a review of youth using online mental health services, 90% reported satisfaction and 86% said they would continue using the service, yet only 65% found the programs helpful ². High engagement metrics are not the same as clinical usefulness, and marketing dashboards that reward the first two numbers will overstate impact against the third.

Rewrite condition pages so that symptom descriptions link to a clinician assessment rather than a self-scoring result. Route quiz completions to a scheduled call, not a score badge. Treat social content as an awareness layer that hands off to trained intake, not as a diagnostic surface.

The FDA’s Digital Health Advisory Committee has been direct about what AI-enabled mental health tools must communicate to users. Transparency should include clear indication that the AI system is not a human therapist, alongside intended use statements, overuse warnings, crisis information, and data use and privacy details ⁴. Every one of those elements has a marketing surface: homepage hero copy, app store descriptions, chatbot welcome screens, and ad creative.

The operational implications are specific. A conversational triage bot that greets a visitor with “How can I help you today?” without identifying itself as automated fails the disclosure standard on the first line. Homepage claims that a tool “provides therapy” or “delivers CBT” without naming the licensed clinicians in the loop misrepresent intended use. Crisis routing that appears only in a footer disclaimer, rather than in the bot’s opening exchange, buries the warning the committee flagged as essential.

Marketing teams should audit every AI-touched surface for four elements in plain language: automated identity, what the tool is for, what it is not for, and how to reach a human in a crisis. The same audit applies to any vendor tool a growth team layers onto intake, from symptom checkers to appointment-scheduling assistants.