Outpatient physical therapy clinic networks Marketing Research 2026

Executive read

The market read

Marketing performance has to connect local media and search with referral relationships, appointments, visits, payer mix, and clinic capacity. Lead volume alone is a weak proxy for business value.

Four lenses translate the public evidence into the questions a category marketer needs to answer: how demand is created, where conversion happens, where measurement breaks, and what to change.

Demand model

Referrals and direct access coexist

Physician relationships, payer rules, local search, and consumer education can all create a first visit at the same clinic.

Conversion moment

Booked care is only the first economic step

Visits, plan of care, payer mix, and clinic capacity determine value after an appointment is scheduled.

Measurement break

Large estates magnify naming drift

A small inconsistency in clinic, brand, or referral-source IDs becomes a material reporting problem across thousands of locations.

Marketer move

Connect acquisition to completed care

Measure by clinic and referral source through booked visit, attended visit, plan of care, and downstream revenue.

Directional conclusions from 5 completed public-source reviews across 10 operators. The remaining company deep dives can change the segment read.

Category evidence

What the existing audits actually support

Four findings worth carrying into the operating model

The existing corpus is deep, but NovaCare and Select Physical Therapy describe overlapping views of Select Medical and must not be double-counted. Counts below keep their original unit and date; no parent/child brands, stores, clinics, laboratories, beds, or partner sites are silently added together.

Terminology used in this edition PT clinicsoutpatient rehabilitationphysical therapy networks

Entity hierarchy

The old 8,529 total double-counted one platform

NovaCare and Select Physical Therapy were both counted at roughly the size of Select Medical's 1,912-clinic outpatient division. The public edition now shows independent systems instead.

Demand

Referral and direct access coexist

Physicians, employers, health plans, and direct consumer demand can create the same evaluation, so referral source cannot be forced into a paid-channel taxonomy.

Architecture

One brand and 100+ brands both exist

ATI runs 850+ clinics under one consumer identity; the U.S. Physical Therapy audit found a federated estate with many consumer domains and analytics containers.

Outcome depth

A booked evaluation is not the economic result

Attendance, authorization, plan of care, completed visits, payer adjustments, discharge, and collected revenue determine value after booking.

Measurement path

The category-specific data contract

The useful outcome sits beyond the lead

Revenue accrues across repeated visits, referrals and payers remain acquisition sources, and capacity is therapist time plus authorized visits rather than a chair or treatment slot.

01 physician, employer, payer, or consumer source
02 clinic and service line
03 request or booking
04 attended evaluation
05 authorization and plan of care
06 completed visits
07 discharge and outcome
08 allowed and collected revenue

Measurement grain

clinic × referral source × service × therapist/capacity × payer

Primary outcome

completed plan-of-care value, not a booked evaluation

Metrics that survive

attended-evaluation CAC
authorization rate
visits per plan
therapist utilization
collected revenue

Creative evidence

The category becomes legible when the real work is visible

These attributed first-party examples show how operators frame need, trust, access, and outcomes. They are editorial evidence, not a performance ranking.

ATI Physical Therapy Website

Thirty years becomes a trust asset, not a corporate milestone

The anniversary frame packages longevity as evidence for a patient deciding whether the network can get them back to movement.

ATI Physical Therapy Website

The therapist team is the brand proof

Faces, uniforms, and a champion metaphor make a large network feel like a coordinated care team rather than a directory of clinics.

ATI Physical Therapy Website

A condition becomes the landing-page entry point

The treatment image matches high-intent condition discovery while leaving room for the location and therapist to carry the conversion.

3 of 3 examples shown

AI adoption

Marketing challenges & AI adoption

Nine workflows: what AI changes, and what it still cannot fix

Outpatient PT networks have AI at the edges of acquisition—ad platforms, SEO automation, and conversation-intelligence vendors—while capacity, referral status, clinic identity, booking outcomes, and privacy controls still require governed joins and human decisions.

Evidence boundary

This guide reviewed ATI Physical Therapy, U.S. Physical Therapy, Select Medical's outpatient brands, and Upstream Rehabilitation through 2026-08-04. A live BrightEdge or Invoca tag proves a deployed tool, not that every AI module is enabled. Google platform capability is labeled as a cross-industry proxy, not PT operator adoption.

Verified industry evidenceOperator exampleCross-industry proxyImprovado hypothesis

Workflow AI today / still manual Evidence Main blocker Practical next use

Creative production

Rare Maturity basis · Cross-industry proxy

AI today

No audited PT operator publicly disclosed generative-AI creative production. Google Ads can generate image, text, and video assets, but operator activation is not visible from public campaigns.

Still manual / non-AI

Teams still choose the clinic, service, offer, patient-safe imagery, claims, and local applicability, then approve every production variant.

Example evidence · Cross-industry proxy Google Ads generative assets ↗

Google documents generative asset capability and advertiser review requirements; it publishes no PT-specific utilization denominator.

Blocker

Clinical claims, patient-image consent, local service availability, and absent asset-to-clinic outcome lineage make unrestricted generation unsafe.

Next use case

Generate controlled format and background variants from approved assets while preserving creative, clinic, service, and approval IDs through booking reporting.

Message and copy generation

Rare Maturity basis · Operator example

AI today

No audited PT operator disclosed AI-written advertising or clinical copy. ATI's named agency case instead describes human-built message themes and value-proposition testing.

Still manual / non-AI

Clinical accuracy, referral and direct-access language, local services, capacity, payer context, and brand approval remain human work.

Example evidence · Operator example ATI Physical Therapy message testing ↗

Cardinal documents patient-centric messaging and value-proposition tests for ATI; the case does not claim AI generation.

Blocker

An unconstrained model can invent outcomes, eligibility, referral requirements, or services that differ by state and clinic.

Next use case

Draft local ads and landing-page modules only from approved service, clinic, direct-access, and claim records, with named publish approval.

Paid-media optimization

Widespread Maturity basis · Cross-industry proxy

AI today

Google Performance Max and Smart Bidding automate bids, budgets, audiences, creative combinations, and attribution. ATI publicly documents Google Ads, but its specific AI campaign settings are not disclosed.

Still manual / non-AI

Operators still decide clinic priority, radius, service, budget, exclusions, capacity status, referral mix, and which conversion signal is trustworthy.

Example evidence · Cross-industry proxy Google Performance Max ↗

Google documents AI across bidding, budget, audiences, creative, and attribution; this is platform capability, not proof of PT activation.

Blocker

A lead or call signal does not reveal referral status, appointment, attendance, reimbursable care, or whether the clinic had capacity.

Next use case

Return validated appointment and attended-care signals by clinic, service, and capacity band before applying value rules or automated budget shifts.

SEO and GEO

Emerging Maturity basis · Verified industry evidence

AI today

BrightEdge Autopilot was observed on the shared Select Medical outpatient PT web estate. It automates technical and internal-link optimization; no complete PT generative-engine program was verified.

Still manual / non-AI

Clinicians and marketers still own treatment accuracy, provider and location facts, editorial judgment, schema review, and the policy for AI crawlers.

Example evidence · Verified industry evidence Select Medical PT sites × BrightEdge Autopilot ↗

Live NovaCare and Select Physical Therapy source showed BrightEdge Autopilot in July 2026; one parent estate is not a category rate.

Blocker

Automation can scale stale clinic facts or weak medical content, while operator crawler policy may intentionally limit AI reuse.

Next use case

Automate technical fixes and internal links first, then monitor citations and route factual conflicts to the location or clinical owner.

UTM, attribution, and data QA

Rare Maturity basis · Improvado hypothesis

AI today

No audited PT network disclosed AI-based attribution QA. ATI joins Invoca and Google Ads in dashboards, while other estates still expose divergent tags, domains, calls, and booking paths.

Still manual / non-AI

Teams maintain campaign naming, clinic crosswalks, booking IDs, referral status, call dispositions, conversion reconciliation, and exception queues.

Example evidence · Operator example ATI clinic-level measurement ↗

The named case joins Google Ads and Invoca for clinic-level reporting; it describes governed dashboards, not autonomous AI QA.

Blocker

AI cannot recover a click, campaign, clinic, or appointment identifier that was never captured or safely joined.

Next use case

Continuously test ad-to-call-to-booking handoffs and flag missing IDs, duplicate conversions, stale clinic mappings, and impossible state changes.

Call analysis

Emerging Maturity basis · Operator example

AI today

ATI currently deploys Invoca and has a named conversation-intelligence use case for call attribution and clinic or hub analysis. Public evidence does not identify every current Signal AI model enabled.

Still manual / non-AI

Teams still define appointment and qualification labels, audit errors, coach staff, manage staffing, and reconcile calls with scheduled care.

Example evidence · Operator example ATI Physical Therapy × Invoca ↗

Invoca documents ATI's use for return on ad spend and operational visibility; current feature configuration is not public.

Blocker

Recorded calls may contain PHI, and an incorrect appointment label can corrupt campaign bidding or staffing decisions.

Next use case

Pilot booking and intent classification against a human-labeled sample, with confidence thresholds, BAA coverage, and reversible ad feedback.

Lead routing and CRM

Rare Maturity basis · Improvado hypothesis

AI today

No audited PT operator disclosed AI lead routing. ATI's public Invoca example uses clinic and hub lookup logic, which is deterministic routing rather than verified AI decisioning.

Still manual / non-AI

Staff still resolve clinic eligibility, direct-access rules, insurance, urgency, staffing, duplicate patients, and referral handoffs.

Example evidence · Operator example ATI hub-and-spoke routing context ↗

Invoca describes lookup-table classification across ATI clinics and hubs; it should not be relabeled as autonomous AI routing.

Blocker

A model cannot route safely without current clinic, capacity, state-rule, payer, and escalation data.

Next use case

Recommend clinic and queue from intent and availability, but enforce deterministic eligibility gates and human review for ambiguous cases.

Scheduling and patient engagement

Rare Maturity basis · Improvado hypothesis

AI today

Audited PT operators expose first-party, location-keyed appointment flows, but no operator-level AI scheduling agent was verified.

Still manual / non-AI

Referral and direct-access questions, payer checks, urgency, slot selection, reminders, missed calls, and clinical escalation remain staff workflows.

Example evidence · Verified industry evidence ATI first-party appointment flow ↗

ATI routes requests with clinic identifiers; the live scheduling surface proves structured booking, not AI engagement.

Blocker

Scheduling depends on referral, insurance, clinician, location, and real capacity—not just an open calendar slot.

Next use case

Start with missed-call recovery, reminders, and constrained rescheduling, with no medical advice and a staffed escalation path.

Compliance and privacy

Rare Maturity basis · Improvado hypothesis

AI today

PT estates use consent and analytics controls unevenly, but no audited operator disclosed AI as the compliance decision-maker. AI can assist inventory, redaction, and exception review only.

Still manual / non-AI

Covered-entity scope, BAAs, consent, state privacy rules, minimum-necessary use, tag approval, and incident ownership remain accountable human decisions.

Example evidence · Cross-industry proxy HHS online tracking guidance ↗

HHS defines the regulated disclosure boundary; it does not make AI a substitute for legal and privacy governance.

Blocker

A model cannot apply policy until surface, entity, data class, purpose, vendor contract, consent state, and destination are known.

Next use case

Automate tag and outbound-field inventory, flag policy or BAA gaps, and require a named approval before collection or activation changes.

Operator map

Operating-model map

Where physical therapy operators sit

Highlighted companies have enough dated public evidence for both directional scores. The full cohort stays in the background for market context.

Paid demand × operating model

Where public demand intensity meets marketing-operations centralization

37 evidence-scored companies · updated August 2026

Quiet consolidatorsCentralized · lower paid intensity Command centersCentralized · higher paid intensity Fragmented foundationsDistributed · lower paid intensity Local firepowerDistributed · higher paid intensity AD Aspen Dental HD Heartland Dental PH PDS Health SD SALT Dental Partners SB Smile Brands SD Smile Doctors SH Sonrava Health FD Forefront Dermatology QP QualDerm Partners UF US Fertility L LaserAway ML Milan Laser SL SEV Laser 4Y 4Ever Young S SkinSpirit VM VIO Med Spa O OVME E Ever/Body B Beltone H HearingLife M Miracle-Ear TJ The Joint AP ATI Physical Therapy N NovaCare SP Select Physical Therapy UP U.S. Physical Therapy UR Upstream Rehabilitation AF American Family Care C CityMD C Concentra UV USA Vein Clinics B Banfield TP Thrive Pet Healthcare VA VCA Animal Hospitals V VetCor M MyEyeDr NV National Vision

Paid demand intensity →

Marketing ops centralization →

Directional scores synthesize dated public ad-library, website, tag-layer, and operating-model evidence. They are not spend, revenue, or vendor-performance scores, and not published rubric composites; positions are directional synthesis.

Aspen Dental: paid demand 70 out of 100; marketing operations centralization 88 out of 100.
Heartland Dental: paid demand 45 out of 100; marketing operations centralization 22 out of 100.
PDS Health: paid demand 72 out of 100; marketing operations centralization 92 out of 100.
SALT Dental Partners: paid demand 22 out of 100; marketing operations centralization 15 out of 100.
Smile Brands: paid demand 68 out of 100; marketing operations centralization 20 out of 100.
Smile Doctors: paid demand 72 out of 100; marketing operations centralization 62 out of 100.
Sonrava Health: paid demand 68 out of 100; marketing operations centralization 34 out of 100.
Forefront Dermatology: paid demand 26 out of 100; marketing operations centralization 86 out of 100.
QualDerm Partners: paid demand 30 out of 100; marketing operations centralization 58 out of 100.
US Fertility: paid demand 55 out of 100; marketing operations centralization 16 out of 100.
LaserAway: paid demand 78 out of 100; marketing operations centralization 70 out of 100.
Milan Laser: paid demand 90 out of 100; marketing operations centralization 74 out of 100.
SEV Laser: paid demand 82 out of 100; marketing operations centralization 72 out of 100.
4Ever Young: paid demand 76 out of 100; marketing operations centralization 55 out of 100.
SkinSpirit: paid demand 68 out of 100; marketing operations centralization 84 out of 100.
VIO Med Spa: paid demand 72 out of 100; marketing operations centralization 55 out of 100.
OVME: paid demand 63 out of 100; marketing operations centralization 80 out of 100.
Ever/Body: paid demand 55 out of 100; marketing operations centralization 78 out of 100.
Beltone: paid demand 30 out of 100; marketing operations centralization 12 out of 100.
HearingLife: paid demand 30 out of 100; marketing operations centralization 84 out of 100.
Miracle-Ear: paid demand 40 out of 100; marketing operations centralization 45 out of 100.
The Joint: paid demand 84 out of 100; marketing operations centralization 52 out of 100.
ATI Physical Therapy: paid demand 30 out of 100; marketing operations centralization 70 out of 100.
NovaCare: paid demand 10 out of 100; marketing operations centralization 74 out of 100.
Select Physical Therapy: paid demand 12 out of 100; marketing operations centralization 66 out of 100.
U.S. Physical Therapy: paid demand 10 out of 100; marketing operations centralization 12 out of 100.
Upstream Rehabilitation: paid demand 14 out of 100; marketing operations centralization 38 out of 100.
American Family Care: paid demand 62 out of 100; marketing operations centralization 15 out of 100.
CityMD: paid demand 40 out of 100; marketing operations centralization 80 out of 100.
Concentra: paid demand 6 out of 100; marketing operations centralization 72 out of 100.
USA Vein Clinics: paid demand 64 out of 100; marketing operations centralization 84 out of 100.
Banfield: paid demand 52 out of 100; marketing operations centralization 76 out of 100.
Thrive Pet Healthcare: paid demand 58 out of 100; marketing operations centralization 30 out of 100.
VCA Animal Hospitals: paid demand 50 out of 100; marketing operations centralization 72 out of 100.
VetCor: paid demand 18 out of 100; marketing operations centralization 38 out of 100.
MyEyeDr: paid demand 85 out of 100; marketing operations centralization 85 out of 100.
National Vision: paid demand 45 out of 100; marketing operations centralization 68 out of 100.

Footprints

Market structure

Largest known operator footprints

Location counts come from the dated research registry and first-party public directories. They are shown to explain operating scale, not to rank quality or performance.

Select Medical outpatient rehabilitation deep dive complete

1,912 clinics

Upstream Rehabilitation deep dive complete

1,200+ clinics

ATI Physical Therapy deep dive complete

850+ clinics

U.S. Physical Therapy deep dive complete

783 owned or managed clinics

FYZICAL Therapy & Balance Centers in research queue

600

Implications

What changes for marketing teams

The measurement design follows the operating model

Model location explicitly

Media, calls, forms, appointments, and revenue need one durable facility identifier.

Separate collection from activation

Privacy-safe collection does not by itself create a governed reporting or activation layer.

Preserve local context

National rollups stay useful only when teams can drill into brand, market, service, and location.

This section describes data-design implications from the research. It is not a claim that every operator has the same stack, privacy obligations, or level of centralization.

Method

Method and boundaries

A dated public-source edition

The segment inherits the parent report method: location directories, sitemaps, booking paths, public web tags, ad transparency libraries, ownership announcements, and public operating-model evidence.

Private CRM history, sales calls, contacts, customer data, internal scoring, and recommendations are excluded. Technology detection means a signal was visible on a reviewed surface; it does not prove enterprise-wide deployment.

Physical therapy edition limits

Five reports represent four independent systems because two consumer brands share one parent platform.
Public web evidence cannot reveal authorization, outcomes, payer adjustments, or collected revenue.
Dated ad-library counts are not spend or performance benchmarks.

4 primary sources in this category synthesis

Compare all thirteen category methods ↗

Marketing performance has to connect local media and search with referral relationships, appointments, visits, payer mix, and clinic capacity. Lead volume alone is a weak proxy for business value.

Referrals and direct access coexist

Booked care is only the first economic step

Large estates magnify naming drift

Connect acquisition to completed care

Four findings worth carrying into the operating model

The old 8,529 total double-counted one platform

Referral and direct access coexist

One brand and 100+ brands both exist

A booked evaluation is not the economic result

The useful outcome sits beyond the lead

clinic × referral source × service × therapist/capacity × payer

completed plan-of-care value, not a booked evaluation

The category becomes legible when the real work is visible

Thirty years becomes a trust asset, not a corporate milestone

The therapist team is the brand proof

A condition becomes the landing-page entry point

Nine workflows: what AI changes, and what it still cannot fix

Creative production

Message and copy generation

Paid-media optimization

SEO and GEO

UTM, attribution, and data QA

Call analysis

Lead routing and CRM

Scheduling and patient engagement

Compliance and privacy

Where physical therapy operators sit

Largest known operator footprints

The measurement design follows the operating model

Model location explicitly

Separate collection from activation

Preserve local context

A dated public-source edition

Compare the operating model