What Is Schema Markup?

Schema markup is code added behind a web page that labels its facts in a standard vocabulary, so a search engine can read "this is the practice name" and "this is the phone number" without guessing. For a health practice, it means your address, providers, specialty and hours are stated in a form machines read directly. Google calls it structured data and says it gives Google explicit clues about the meaning of a page.

How schema markup works

Your visitors read a page as design and sentences. A machine reads the same page as raw text. A line such as "Dr. Lee, ND, Tuesdays in Boulder" is clear to a person. To software it could be a name, a title, a place or a schedule. Markup removes the ambiguity by attaching a label to each fact.

The labels come from schema.org, a shared vocabulary that Google and other search engines read. Google's documentation says most Search structured data uses schema.org vocabulary, and recommends the JSON-LD format in most cases because it is the easiest to implement and maintain. JSON-LD is a block of code placed in the page that lists the facts as labeled pairs, such as name followed by your clinic name. Visitors do not see it, and it can sit separate from the design of the page.

A typical practice has four layers of markup.

• Organization. The practice as a whole, with name, URL, logo, phone and links to official profiles. Google's guide recommends placing it on your home page or a single page that describes the organization.
• Local business. The physical location, with address, hours and phone. Google advises using the most specific business subtype available, and a clinic has health-specific options, covered in medical schema.
• Person or physician. Each provider, with credentials, specialty and the clinic they work at.
• Page-level types. Article markup on blog posts, breadcrumb markup and similar types for specific pages.

Markup carries two limits. First, Google says structured data helps it understand a page and can make a page eligible for richer results, but eligibility depends on following Google's guidelines and including the required properties. Second, Google says there are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations, and that no special schema.org structured data is needed for those features. Treat markup as clean labeling. Other AI tools may read the code too, but I have not seen a vendor confirm a benefit, so I do not treat it as a lever.

Two rules keep markup honest. Mark up only what a visitor can see on the page. Google's guidelines say not to mark up content that is not visible to readers, and that structured data must be a true representation of the page content. And keep the set short. For most practices the useful types are the business, each provider as a person, each service and articles. Every extra type is one more thing to keep accurate.

What the evidence says about AI citations is mixed at best. SE Ranking analyzed 129,000 domains and 216,524 pages and reported by Search Engine Journal in November 2025 that referring domains, meaning other sites linking to you, were the strongest predictor of ChatGPT citations. Pages with FAQ schema averaged fewer citations than pages without it, 3.6 versus 4.2. That is an average across the sample and correlation, and it does not show that FAQ markup causes fewer citations. It gives no support to stamping it on every page.

Google also changed what FAQ markup does in search. In August 2023 it limited FAQ rich results to well-known government and health websites, and in 2026 it stopped showing them in Search at all. Older articles that promised expanding question boxes are out of date.

Markup also connects to identity. The markup on your entity home page is where Google's url and sameAs fields live, which is how you point to your other official profiles and help Google recognize the same practice across the web.

Example: schema markup at a health practice

This example is illustrative, not a transcript. A naturopathic clinic has two locations, one in Boulder and one in Fort Collins. The website lists both addresses in a footer, and each location page repeats its own phone number and hours as text. A plugin added the clinic's name and one address as markup, taken from the first location entered during setup.

A system reading the markup sees one clinic with one address. It sees two phone numbers in the page text and cannot tell which belongs where. If the markup is wrong, the system must either trust the code or the text, and nothing on the page tells it which to prefer. A wrong value can also be repeated back to a patient, which is an AI brand accuracy problem.

The fix is to give each location page its own local business markup. The Boulder page labels the Boulder address, phone and hours. The Fort Collins page labels its own. The home page carries organization markup with the clinic name, URL, logo and links to the official profiles. Every labeled fact matches what a visitor can see on that page.

Real data helps set expectations. Google's own documentation says markup can make a page eligible for richer results and states no guarantee. It also says no special markup is needed for its AI features. SE Ranking's data showed no citation advantage for FAQ markup. The value of markup is accuracy, since a labeled fact is easier for a machine to read than an inferred one, and that value shows up when the labels are correct and match the visible page.

For testing, I recommend the Schema Markup Validator at validator.schema.org, run under the schema.org project. Google's Rich Results Test checks markup against Google's rich result features, so it answers a narrower question than whether your markup is correct. For health types such as MedicalClinic and Physician, the Schema Markup Validator is the better first check.

How to check yours in 60 seconds

1. Go to the Schema Markup Validator at validator.schema.org.
2. Enter your home page address and run the test.
3. Read the detected types, such as Organization, LocalBusiness or MedicalClinic, and open each one to see its fields.
4. Compare the name, address, phone and URL in the results to what a visitor sees on the page.

Nothing detected means the page has no schema.org markup. A detected type with an old address or phone number means the markup is stale, which needs fixing before anything else. Repeat the test on one provider page and one location page, because plugins often cover only the home page.

What to do about it

1. List the facts to label. Start with practice name, address, phone, URL, logo and links to your official profiles, and write each one down exactly as it appears on your Google listing.
2. Choose the right types. See medical schema for the health-specific options.
3. Add the markup as JSON-LD on your entity home page, and add provider details on each provider page.
4. Make every labeled fact match what is visible on the page.
5. Test each page in the Schema Markup Validator, fix errors and read the warnings.
6. Retest after every change, and recheck when providers, hours or addresses change. Keep a short change log with the date and what changed.
7. Ask your web developer or plugin vendor which fields the markup fills automatically, and where the values come from. If a plugin pulls the address from a settings page, update that page first, then the markup follows.
8. Skip markup types you cannot support with visible content. Label only the services you offer and the providers you employ.