Healthcare category deep dive Industry research 01.07 · August 2026

Eye care and optical retail groups: Marketing Category Deep Dive

National Vision and MyEyeDr show the category's core collision: clinical exams and retail transactions share locations but use different clocks and economics. Written for marketing leaders, Marketing Operations, and analysts who need the category's real measurement path.

Evidence base
8 operators
Company deep dives
2 complete
Creative examples
3 attributed
Evidence status
2 company audits complete
01

Executive read

The market read

The operating model has to reconcile retail sales and clinical appointments without losing the location, product, payer, and promotional context that explains performance.

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

Retail and clinical journeys overlap

Eyewear discovery, exam booking, insurance eligibility, and e-commerce can start in different channels and converge at one store.

Conversion moment

An exam and a product sale are separate events

Appointment value, frame or lens purchase, payer reimbursement, and repeat purchase operate on different clocks.

Measurement break

Product context is lost in clinic reporting

A store can appear efficient on appointment volume while promotion mix or product margin tells the opposite story.

Marketer move

Keep store, patient, product, and payer together

Use one identity and location model across online discovery, exam booking, insurance, and retail transactions.

Directional conclusions from 2 completed public-source reviews across 8 operators. The remaining company deep dives can change the segment read.

02

Category evidence

What the existing audits actually support

Four findings worth carrying into the operating model

National Vision and MyEyeDr show the category's core collision: clinical exams and retail transactions share locations but use different clocks and economics. 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 optometry groupseye care networksoptical retailers
Denominator

Store count and exam-enabled count are not interchangeable

Warby Parker ended 2025 with 323 stores, 285 offering in-person exams. National Vision reported 1,274 stores; MyEyeDr and Visionworks each report 700+ estates.

Two outcomes

An exam and a product sale are separate conversions

Prescription, frame or lens order, fulfillment, return, and repeat contact purchase must remain distinct from an attended exam.

Economics

Payer and product mix change store performance

Managed-care strength, self-pay traffic, promotion, product margin, inventory, and returns can tell a different story from appointment volume.

Architecture

Central brand and local-practice federation both scale

MyEyeDr uses one consumer identity across 900+ practices; AEG preserves local optometry brands across 500+ practices.

03

Measurement path

The category-specific data contract

The useful outcome sits beyond the lead

A product can be purchased without a clinical appointment. Inventory, returns, managed-care reimbursement, e-commerce, and store conversion make this a retail-health system rather than a clinic-only funnel.

  1. 01 channel, query, or offer
  2. 02 browse or insurance check
  3. 03 exam booking or direct commerce
  4. 04 exam and prescription
  5. 05 eyewear or contact order
  6. 06 fulfillment or return
  7. 07 margin and repeat purchase
Measurement grain

store × clinician × service × payer × product/SKU × promotion

Primary outcome

exam and retained product margin, measured separately

Metrics that survive
  • attended-exam CAC
  • exam-to-order
  • gross margin
  • return rate
  • repeat contacts purchase
04

Creative evidence

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.

05

AI adoption

Marketing challenges & AI adoption

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

Eye-care and optical groups have a clear live AI paid-media example and strong signals in creative and call intelligence, but practice economics, exam capacity, booking feedback, local facts, and privacy controls still determine whether automation improves patient acquisition.

Evidence boundary

This guide reviewed MyEyeDr, National Vision, and AEG Vision evidence through 2026-08-04. MyEyeDr's live `_AI` filename is an AI-production signal, not independent proof of its generation method. AEG's PMax use is a named operator case; Invoca presence does not prove that every AI call feature is active.

Verified industry evidenceOperator exampleCross-industry proxyImprovado hypothesis
Workflow AI today / still manual Evidence Main blocker Practical next use

Creative production

Emerging Maturity basis · Verified industry evidence
AI today

MyEyeDr's current homepage serves a production image whose filename includes `_AI`. That verifies an AI-labeled asset in use, but not the model, workflow, or degree of generation.

Still manual / non-AI

Teams still own concepts, model releases, patient-safe representation, offers, brand fit, local applicability, and final approval.

Example evidence · Verified industry evidence MyEyeDr AI-labeled homepage asset ↗

The asset was live on the homepage and carried a June 2026 production path when checked 2026-08-04; filename evidence is not full provenance.

Blocker

A filename cannot prove source rights, model release, brand safety, or whether depicted people and settings are real.

Next use case

Require source metadata and approval status, then generate controlled format or background variants tied to campaign, market, and booking outcomes.

Message and copy generation

Rare Maturity basis · Cross-industry proxy
AI today

No audited vision operator publicly disclosed AI copy generation. Google can create campaign text when enabled, while AEG's named creative case describes human strategy and message development.

Still manual / non-AI

Teams still define symptoms, offers, insurance language, exam value, local service facts, tone, and clinical or legal review.

Example evidence · Operator example AEG Vision performance creative ↗

The named case describes story-driven concepts and human creative inputs to PMax; it does not claim AI-written copy.

Blocker

Generated symptom or benefit claims can overstate clinical meaning, and offers vary by brand, payer, market, and practice.

Next use case

Draft from approved exam, service, offer, insurance, and location facts while keeping a human publish gate and source references.

Paid-media optimization

Emerging Maturity basis · Operator example
AI today

AEG Vision publicly documents a live Performance Max program. National Vision separately describes search and social algorithms relearning from booking signals after its 2026 site replatform.

Still manual / non-AI

Operators still decide creative inputs, market groups, practice capacity, margin, show rate, new-patient value, budgets, and exclusions.

Example evidence · Operator example AEG Vision × Performance Max ↗

The April 2026 named case documents PMax using creative and conversion signals; the reported result is an agency case, not an audited category benchmark.

Blocker

An exam-booking signal can be lost during a site change or can value practices equally despite different slots, show rates, tickets, and margins.

Next use case

Feed validated new-patient, attended-exam, capacity, and value bands into bidding while retaining market and practice-level guardrails.

SEO and GEO

Rare Maturity basis · Improvado hypothesis
AI today

No current operator AI-SEO deployment was verified. National Vision describes its new platform as preparation for future AI consumer marketing and agentic commerce, not a live adoption claim.

Still manual / non-AI

Location and provider facts, exam availability, schema, migration QA, editorial content, listings, and local differentiation remain operational work.

Example evidence · Operator example National Vision future-ready commerce platform ↗

The May 2026 earnings call frames agentic commerce and AI personalization as future readiness, while documenting current search-signal recovery work.

Blocker

AI answers and local search both fail when location, provider, service, insurance, and booking facts disagree across systems.

Next use case

Create a governed location-provider-service graph, monitor search and answer-engine citations, and route factual conflicts to the operating owner.

UTM, attribution, and data QA

Emerging Maturity basis · Operator example
AI today

AEG combines portfolio-wide UTMs, server-side tracking, direct matchback, call outcomes, and PMax attribution. The operator describes a governed foundation, not autonomous AI QA.

Still manual / non-AI

Teams maintain practice mappings, new-versus-existing patient logic, call and booking joins, finance inputs, and campaign exception review.

Example evidence · Operator example AEG Vision measurement rebuild ↗

AEG leaders describe UTMs, call tracking, direct matchback, and campaign feedback across a complex practice estate.

Blocker

Hundreds of domains and campaigns cannot share learning when practice, patient type, call, appointment, and financial outcomes lack governed keys.

Next use case

Add automated tests for broken UTMs, missing booking callbacks, stale practice mappings, duplicate conversions, and sudden signal loss after releases.

Call analysis

Emerging Maturity basis · Operator example
AI today

AEG uses Invoca and passes booked-appointment call outcomes into measurement. Public operator testimony does not specify which current AI classification, summary, or QA modules are enabled.

Still manual / non-AI

The central call center still handles conversations, validates outcomes, coaches agents, resolves exceptions, and connects calls to attended exams.

Example evidence · Operator example AEG Vision × Invoca ↗

AEG names Invoca as its call-tracking partner and describes booked-appointment feedback; AI feature configuration remains private.

Blocker

Transcripts can contain health and insurance context, while a false booking label can distort media optimization and coaching.

Next use case

Validate intent, booking, and quality labels against human-reviewed calls before using them for campaign feedback or automated QA.

Lead routing and CRM

Rare Maturity basis · Improvado hypothesis
AI today

No audited operator disclosed AI lead routing. AEG moved advertising and website calls to a central call center and described a homegrown CRM, which are operational and deterministic choices.

Still manual / non-AI

Staff still resolve location, insurance, exam type, urgency, available slots, duplicate patients, and practice exceptions.

Example evidence · Operator example AEG centralized call routing ↗

AEG leaders explain the centralized call-center decision and homegrown CRM; the source does not call the routing AI-driven.

Blocker

Routing requires current slot, practice, payer, service, and patient-status data before model recommendations are safe.

Next use case

Recommend destination and priority from intent and capacity while deterministic eligibility rules and staffed escalation retain control.

Scheduling and patient engagement

Rare Maturity basis · Operator example
AI today

No current appointment agent was verified. National Vision documents high online exam-booking importance and future agentic-commerce readiness; AEG currently relies on online slots and a human call center.

Still manual / non-AI

Exam type, insurance, provider and slot selection, rescheduling, reminders, no-show recovery, and clinical escalation remain staff-owned.

Example evidence · Operator example National Vision online exam journey ↗

The May 2026 earnings call connects search and social signals to online exam booking but describes agentic commerce as future-facing.

Blocker

A safe agent needs accurate exam inventory, insurance and patient-status rules, and an escalation path for clinical questions.

Next use case

Start with constrained slot finding, reminders, and rescheduling, recording source and outcome while handing clinical or coverage exceptions to staff.

Compliance and privacy

Rare Maturity basis · Improvado hypothesis
AI today

Operators disclose consent, server-side tracking, or privacy-management controls, but none in the audited set disclosed AI as the compliance authority.

Still manual / non-AI

Entity scope, state rules, BAAs, sensitive-health targeting, consent, tag approval, patient communications, and incident ownership remain human decisions.

Example evidence · Cross-industry proxy Google health-sensitive targeting policy ↗

Google restricts personalization based on personal health content; platform optimization does not override audience or data-use policy.

Blocker

State, entity, surface, purpose, consent, data class, and vendor contract must be known before a policy check is meaningful.

Next use case

Automate surface and tag inventory, detect changes or policy gaps, and require named legal or privacy approval before activation.

06

Operator map

Operating-model map

Where eye care & optical 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
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.
07

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.

National Vision deep dive complete
1,274
MyEyeDr deep dive complete
900+ practices
AEG Vision in research queue
500+
Visionworks in research queue
700+ locations
Warby Parker in research queue
323 stores; 285 exam-enabled
San Antonio Eye Center in research queue
16
08

Implications

What changes for marketing teams

The measurement design follows the operating model

01

Model location explicitly

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

02

Separate collection from activation

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

03

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.

09

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.

Eye care & optical edition limits
  • Stores, practices, clinics, and exam-enabled stores are different footprint units.
  • Public appointment surfaces cannot expose payer adjudication, inventory, returns, or margin.
  • The two complete audits do not support category-wide media-intensity claims.
4 primary sources in this category synthesis
Compare all thirteen category methods ↗