Operations
Part of Getting Healthcare SEO Right the First Time
How to Scrutinize Healthcare SEO Case Studies Before Citing Them
This guide explains how to read healthcare SEO case studies, what a usable one contains, and how to audit vendor claims before citing them in healthcare marketing.
What to take away
- Checkable healthcare SEO case studies exist, but they are uneven. Cleveland Clinic, Mayo Clinic, Geisinger, UPMC and several NHS trusts all publish or present their own search and content work, and each discloses a different amount of method.
- Read the method before the result. If the case does not say what changed and when, the number means nothing.
- Ranking gains are not the outcome. Appointment requests, completed intake and fewer wrong-service calls are.
- Objective claims need a reasonable basis before they run, per the FTC substantiation policy and the FTC's Health Products Compliance Guidance on health-related claims.
- Reviews used as proof must be genuine and unconnected to incentives, per FTC guidance for marketers.
Healthcare SEO case studies get cited in pitches far more often than they get read. The five organizations below are real and their public pages are easy to reach, but how much method each one publishes varies from one to the next — and that is the part worth checking.
Each entry names the organization, the public change it made and where the account of it lives. A reviewer can check all three before repeating the approach on another site.
Five cases you will see cited
| Case | What changed | Reported result | Where the write-up lives |
|---|---|---|---|
| Cleveland Clinic | Health library restructured around conditions, not departments; articles carry a clinician review note and a date | Organic sessions and page-level engagement; no baseline or window published | No primary write-up; third-party summaries only. Live content at my.clevelandclinic.org/health |
| Mayo Clinic | Symptom and disease pages rebuilt for query intent, in one repeated section order: symptoms, causes, risk factors, diagnosis, treatment | Page-level engagement; no appointment data published | No primary write-up; conference talks and the live templates at mayoclinic.org/diseases-conditions |
| Geisinger | Service-line pages tied to appointment requests | Appointment requests named as the outcome; baseline and window not published | Health system marketing presentations; geisinger.org |
| UPMC | Local and provider pages consolidated after duplication | Organic traffic to the consolidated pages; window not stated | Conference sessions and trade press; upmc.com |
| NHS trusts | Appointment and service pages rewritten for plain language, under the NHS digital service manual style guide | Service-page performance in trust reports; most give no baseline | Trust digital reports and NHS England guidance; service-manual.nhs.uk |
Cleveland Clinic's library work is usually reported against organic sessions and page-level engagement, not appointments. That is a real limit, and any summary that skips it is one you should not repeat.
When you read or commission a service-line case, ask for a 6-to-12-month window measured against appointment requests, with the quarter before launch as the baseline. Without those two details, a percentage cannot be compared with anything — and the organizations below rarely publish both.
Cleveland Clinic rebuilt its public health library around conditions rather than departments. Articles at my.clevelandclinic.org/health carry a clinician review note and a date, so the method is on the page. Mayo Clinic's disease pages at mayoclinic.org repeat one section order and carry a staff byline; the template is the change.
Geisinger, based in Danville, Pennsylvania, describes service-line pages built around appointment requests in its marketing presentations. UPMC, based in Pittsburgh, keeps a single Find a Doctor directory at upmc.com. NHS trusts publish their own content rules in the digital service manual.
What a usable case contains
A case you can act on names the site, the service line, the technical state before the change, the content owner, and the concurrent changes that might explain the result. Without that list, a ranking jump is a coincidence with a chart.
Ask for the Search Console query export, the GA4 conversion definition and a crawl of the site from before and after, from a tool such as Screaming Frog. Those three files settle more arguments than a slide deck does.
CDC's Step 2: Describe the Program asks for exactly this: a shared description and a logic model. For a health system the resources are clinicians and schedulers, the outputs are condition pages, and the outcome is a completed appointment.
Apply it to a service page. Say how the page moves a patient from search to booked visit, then measure that path.
CDC's Step 3: Focus the Evaluation Questions and Design narrows this to questions that match intended use, program stage and available resources. Decide in advance what result would make you stop, and what would make you scale.
Audit the claim before you cite it
The FTC's advertising substantiation policy requires a reasonable basis for objective claims before they are published, and its Health Products Compliance Guidance applies that standard to health-related claims. A vendor saying a client "tripled organic traffic" is making that kind of claim.
The FTC guidance for marketers using reviews covers fake feedback, selective requests, conditioned incentives and hidden relationships. The same tests apply when a healthcare SEO case study collection leans on provider testimonials.
Separate four things: what the provider claims, what the client says, what the analytics show, and what an outside party verified. Most published cases contain only the first two.
HIPAA limits what a health system can publish about patient behavior. Under the HIPAA Privacy Rule at 45 CFR Part 160 and Part 164, a case can report sessions, form fills and page-level engagement without exposing protected health information. It cannot report that a named patient booked a scan.
That is why appointment requests appear as an aggregate count, if at all, and why a baseline is often given as a range. The vendor contract and the analytics setup decide what may be published. Reproduce one material claim with your own data before you copy the method.
Common questions
Are these real campaigns?
Yes. Cleveland Clinic, Mayo Clinic, Geisinger, UPMC and several NHS trusts have published or presented accounts of their search and content work that a reader can still find. What varies is how much method each one discloses, which is why the table lists where each account lives rather than a single result.
Can a ranking gain prove the work caused it?
No. Rankings move with algorithm updates, competitor changes and seasonality. Read them alongside query context, technical changes, user behavior, appointment records and capacity, and say which alternative explanations you could not rule out.
When does a case justify scaling the approach?
When the evidence is relevant to your service line, the patient path actually completes. Privacy, access, capacity, cost and maintenance guardrails must hold. A case from a 40-provider system rarely transfers to a three-location practice without re-testing.
What should a practice ask a vendor citing a case?
Ask for the baseline period, the change log, the metric definition and the client contact. If the vendor cannot produce the method, treat the case as a lead to test, not a verified result.
