Healthcare Clinic GEO Case Study: 6-Location Specialty Group, +717% Appointment Requests
A 6-location US specialty clinic group (anonymized as “Group D”) in the dermatology + cosmetic dermatology space joined the CapstonAI platform in October 2025. Patient acquisition costs had risen to $310 per booked appointment via Google paid + Meta paid. Compliance constraints (HIPAA, FDA cosmetic claims, state medical board rules) meant the marketing team couldn’t simply replicate D2C tactics. Over 90 days the group executed a compliance-first GEO playbook focused on Physician schema, medical reviewer attribution, and city-level + condition-level content. Result: AI-attributed appointment requests grew from 12/month to 98/month (+717%), cost per appointment dropped 45%, and the group became the most-cited specialty option in their 3 largest metros for 11 of 18 monitored prompts.
Get a free GEO audit for your Healthcare / Specialty Clinic company → Pricing
Company snapshot (anonymized)
| Attribute | Value |
|---|---|
| Industry | Specialty medical clinic group — dermatology + cosmetic dermatology |
| Annual revenue | $18M |
| Employees | 84 (12 physicians, 28 nurses + medical assistants, 18 admin, 26 support) |
| Locations | 6 clinics across 3 metros (West Coast) |
| Patient mix | 62% medical (insurance), 38% cosmetic (cash-pay) |
| Pre-existing channels | Google paid (44%), Meta paid (16%), referral (24%), organic search (12%), other (4%) |
| Setup investment | $22,000 (includes compliance review across all content) |
| Internal owner | Director of Marketing (0.4 FTE) + medical reviewer (0.1 FTE physician time) + compliance officer review at gates |
Starting point — Q4 2025 baseline
| Metric | Value |
|---|---|
| ChatGPT brand citations (panel of 30 condition + city prompts) | 1 |
| Perplexity citations on “best dermatologist [city]” prompts | 3 |
| Gemini citations | 0 |
| AI-attributed appointment requests | 12/month |
| MedicalOrganization schema coverage | 0% |
| Physician schema (per-doctor) | 0 of 12 physicians had Person + Physician schema |
| Medical reviewer attribution on clinical content | 0% of pages |
| FAQPage schema on condition pages | 0% |
| Wikipedia / Wikidata | No Wikipedia. No Wikidata for clinic group or any physician. |
| Google Business Profile | All 6 locations, mixed completeness, 4.4-4.7 star ratings |
| Cost per booked appointment (paid + organic blended) | $310 |
The 90-day playbook executed
- Days 1-5 — Compliance gate + content audit. Compliance officer + outside healthcare counsel reviewed every existing condition page. Identified 23 pages with marketing claims requiring revision (e.g., “best treatment for X” replaced with educational framing). Built a content review checklist that every new page would pass before publication.
- Days 6-20 — Schema deployment. Deployed MedicalOrganization schema sitewide. Built Physician + Person schema for all 12 physicians with NPI numbers, state license numbers, board certifications, medical school, residency, fellowship, languages spoken, sameAs links to state licensing boards and Doximity profiles. Deployed MedicalClinic schema per location with hours, address, phone, accepted insurance.
- Days 15-30 — Medical reviewer attribution rollout. Added “Medically reviewed by [Physician], MD, board-certified in [Specialty], on [Date]” to every clinical page (47 pages total). Each reviewer tag schema-tagged with Person reference. Established a quarterly review cadence so reviewer dates stay current.
- Days 20-45 — Condition + procedure content cluster. Published 28 condition + procedure pages: 14 medical (acne, eczema, psoriasis, skin cancer screening, etc.), 14 cosmetic (Botox, fillers, lasers, microneedling, chemical peels, etc.). Each page: educational tone, no superlative claims, citations to NIH/AAD/peer-reviewed journals, FAQPage schema with 8-12 patient-typical questions, medical reviewer attribution, named physician author.
- Days 25-50 — City + neighborhood pages per location. Built 6 location pages with full LocalBusiness + MedicalClinic schema. For 3 largest metros, built 12 neighborhood-level pages (“dermatologist in [neighborhood]”) with educational + service info. Avoided keyword-stuffing; focused on genuinely useful local information.
- Days 30-60 — Citation density build-out. Every clinical page received 4-12 outbound citations to NIH, AAD, JAMA Dermatology, Mayo Clinic, peer-reviewed studies. Perplexity rewards source-dense health content heavily. Citation count went from average 0.2/page to average 6.8/page across clinical pages.
- Days 35-70 — Wikidata + verified credentials. Built Wikidata entries for the clinic group (24 properties) and for the 4 most-published physicians (board certifications, medical school, residency, fellowship, publications). Wikipedia not pursued (notability bar not met for the group at $18M revenue).
- Days 45-80 — GBP + review depth across 6 locations. Standardized GBP across all 6 locations: identical service categories, accurate hours, weekly posts per location, all Q&A populated, all photos refreshed. Active review push (HIPAA-compliant: simple post-visit text with review link, no PHI). Review counts grew 22-58% per location over 45 days.
- Days 60-90 — Continuous prompt monitoring + reactive updates. Weekly scrape of 30 prompts spanning 18 conditions × 3 metros + brand prompts. Reactive content cycle: when a competitor was cited, build or update equivalent compliance-cleared content within 21 days. 14 reactive moves shipped.
Results — Q4 2025 vs. Q1 2026
| Metric | Q4 2025 | Q1 2026 | Delta |
|---|---|---|---|
| ChatGPT brand citations (panel of 30 prompts) | 1 | 11 | +1,000% |
| Perplexity citations on “best dermatologist [city]” | 3 | 18 | +500% |
| Gemini citations | 0 | 7 | — |
| AI-attributed appointment requests | 12/mo | 98/mo | +717% |
| Cost per booked appointment | $310 | $170 | −45% |
| Telehealth visit conversion from AI traffic | — | 11.4% | — |
| Cosmetic procedure inquiries (cash-pay) from AI | 3/mo | 41/mo | +1,267% |
| Insurance-covered consult requests from AI | 9/mo | 57/mo | +533% |
| Google Business Profile review count (all 6 locations) | 1,847 | 2,612 | +41% |
| Average GBP rating across 6 locations | 4.55 | 4.68 | +0.13 |
| Net incremental revenue Q1 2026 | — | +$317,000 | — |
| Payback on $22k setup | — | 88 days | — |
Lessons learned
- Medical reviewer attribution was the single highest trust-signal lift. Pages with named MD reviewer + date + credentials were cited 4.2x more than identical content without attribution.
- Citation density (NIH, AAD, peer-reviewed) was the structural unlock for Perplexity. Source-dense educational content outperformed marketing-style content 3.8x for citation rate.
- Schema for individual physicians (Person + Physician with NPI + board certs) made specific doctors recommendable. AI engines began naming individual physicians in “who should I see for X in [city]” prompts.
- City + neighborhood pages outperformed single “locations” pages by a wide margin. Patient AI prompts are local (“dermatologist near [neighborhood]”); 18 micro-pages captured 6x the citations of a single locations directory.
- Compliance review was a feature, not a bug. The same edits that satisfied compliance (educational framing, citations, no superlative claims) were also what AI engines preferred. Compliance and GEO aligned cleanly.
What we’d do differently
- Would have built the Wikidata entries in week 1, not week 6. They take 4-7 weeks to fully propagate as identity anchors; starting earlier would have compressed the citation curve.
- Would have invested in 1-2 physician-authored published research pieces (Doximity, JAAD letters, conference abstracts). Original physician research adds Wikipedia-adjacent authority and was the missing piece in our citation profile.
- Would have started GBP review push on day 1 instead of day 45. Reviews compound slowly and the 45-day delay cost ~80 reviews across 6 locations.
FAQ — replicability
Does this replicate at a single-location independent practice?
Yes with budget at $9-14k. Same playbook, smaller surface area. Expect 5-7 months to ROI vs. 88 days at multi-location scale because review and citation counts ramp slower.
What about hospital systems (50+ physicians)?
Same playbook applies but Wikipedia + research-publication track becomes the dominant moat. Major hospital systems with Wikipedia presence (Mayo, Cleveland Clinic, Johns Hopkins) are cited 12-30x more than peers because of Wikipedia anchoring. Setup at hospital scale: $80-200k.
Is patient PHI ever at risk in this work?
No. All GEO content is public-facing educational and structural (schema, content, citations). No PHI is referenced or generated. Standard HIPAA controls remain in place. Compliance officer should review all content gates regardless.
Related reading
- AI Citation Tracking
- How to Rank in Perplexity
- How to Build a Prompt Panel for Tracking
- WordPress AI SEO Plugin
- CapstonAI Platform
Want results like this for your Healthcare / Specialty Clinic company?
Free Healthcare / Specialty Clinic GEO audit →
Last updated: May 2026. Sources: CapstonAI customer cohort Q1 2026 (8 healthcare organizations tracked, this group’s full prompt panel + GA4 + practice-management appointment data with permission, anonymized for publication). Group medical director and compliance officer reviewed and approved this case study. All content claims verified compliant against HIPAA, FDA cosmetic claims rules, and applicable state medical board guidelines.