Schema markup is useful because it gives search engines structured information about the content on a page. It can help Google understand specific facts and make eligible pages available for supported rich-result features such as product information, recipes, events, videos, breadcrumbs and other enhanced search experiences.
What schema markup does not do is just as important. Adding JSON-LD does not automatically improve rankings. It does not guarantee a rich result. And there is no documented universal schema markup that makes ChatGPT, Google AI Mode, Gemini, Perplexity or another AI system cite or recommend your brand.
That distinction has become increasingly important as structured data gets pulled into conversations about generative engine optimization. Schema remains a valuable part of technical SEO. It just needs to be implemented for the reasons the technology actually supports.
The practical rule: use structured data to describe real page content accurately and to support documented search features. Do not use schema as a substitute for useful content, crawlability, site architecture or factual accuracy—and do not promise AI citations because markup exists.
What is schema markup?
Schema markup is structured data that uses a shared vocabulary, most commonly Schema.org, to describe information on a webpage in a standardized way.
Consider a product page. A person reading the page may immediately recognize the product name, brand, price, availability and review rating because those details are presented visually. Structured data can label those facts explicitly so compatible systems do not have to rely only on the visual layout or surrounding prose to determine what each value represents.
The same concept applies to articles, organizations, people, local businesses, events, recipes, videos and many other kinds of information.
Three terms are often used interchangeably even though they mean different things:
| Term | What it means |
|---|---|
| Structured data | A standardized way of describing information so software can interpret its meaning and structure. |
| Schema.org | A shared vocabulary of types and properties used to describe things such as products, organizations, people and articles. |
| JSON-LD | A format that can be used to publish linked structured data on a webpage. Google also supports Microdata and RDFa for structured data. |
Google generally recommends JSON-LD when a site's technical setup allows it because it is usually easier to implement and maintain than markup woven directly through the visible HTML.
What does structured data actually do for SEO?
For Google Search, structured data has two well-established jobs.
First, it provides explicit information about the content and entities represented on a page. Google describes structured data as a standardized format that helps it understand page content and classify the information being presented.
Second, supported structured data can make a page eligible for richer search appearances. Depending on the type of page and Google's current Search features, that can include additional product information, recipe details, event information, breadcrumbs, video features and other enhancements beyond an ordinary text result.
Eligibility is the important word. Correct markup does not guarantee that Google will display a rich result. Search appearance still depends on Google's systems, the query, device, location, page quality and other factors.
Does schema markup improve Google rankings?
Schema markup should not be treated as a direct ranking shortcut.
Its clearest SEO value comes from helping Google understand structured information and enabling supported search features when a page qualifies. A richer result may also change how people interact with a listing, but the presence of markup itself should not be presented as an automatic ranking boost.
This distinction matters because it changes how teams prioritize implementation. There is little value in spending weeks marking up every possible field simply because Schema.org provides a property for it. Start with information and search features that genuinely matter to the business.
Does schema markup improve AI search visibility?
There is no universal rule showing that adding schema markup causes a page to receive more AI citations or causes a brand to be recommended more frequently.
For Google's own generative Search experiences, the guidance is unusually clear. Google says there is no special Schema.org markup required for AI Overviews or AI Mode, and structured data is not an additional requirement for appearing in those experiences. The same core SEO foundations continue to matter.
That does not make structured data irrelevant. If a page contains structured data, it should remain accurate and consistent with the visible content. And if the markup supports a useful Google Search feature, that remains a legitimate reason to implement it.
What marketers should avoid is turning those established uses into an undocumented causal claim such as:
schema → higher AI confidence → more citations.
Different AI products retrieve and use information differently. Their behavior also changes over time. Measure AI visibility as its own outcome rather than assuming a structured-data implementation caused it.
LSEO's guide to measuring AI visibility and citation share explains how to evaluate that performance separately.
Schema.org and Google-supported structured data are not the same thing
Schema.org contains a large vocabulary for describing real-world things and relationships. Google Search supports a smaller set of structured-data implementations for specific Search features.
That difference is important.
A Schema.org type can be completely valid without creating a special Google rich result. Likewise, a property may be useful for representing information even when Google does not require it for a Search feature.
If your objective is a Google search appearance, use Google's current documentation for that specific feature as the implementation standard. Do not assume that every type or property listed on Schema.org produces a corresponding Search enhancement.
| Page or content type | Common structured-data use | What to remember |
|---|---|---|
| Product pages | Product, Offer and related properties | Price, availability and other marked-up facts must accurately reflect the product and offer. |
| Articles | Article, NewsArticle or BlogPosting where appropriate | Authorship, dates, headline and other properties should agree with the visible article. |
| Businesses and organizations | Organization or the appropriate business type | Use accurate identity and business information rather than stuffing unrelated properties into the graph. |
| Author/profile pages | ProfilePage and Person where appropriate | The markup should reflect the person the page is actually about. |
| Video | VideoObject and supported video properties | Use current Google video documentation rather than assuming every Schema.org property changes Search presentation. |
| Site hierarchy | BreadcrumbList | Markup should represent meaningful navigation and hierarchy rather than an invented path built only for search engines. |
What happened to FAQ and HowTo schema?
This is an important example of why schema strategy must follow current platform documentation rather than old SEO checklists.
FAQPage markup still exists in Schema.org, but Google significantly restricted FAQ rich results. They are generally shown only for well-known, authoritative government and health websites. For most commercial sites, adding FAQPage markup should not be justified by an expectation that an expanded FAQ result will appear in Google Search.
Google also retired its HowTo rich-result feature and removed the dedicated HowTo structured-data documentation for that Search appearance.
That does not mean useful FAQ content or instructional content stopped mattering. It means the visible content should exist because it helps people. Structured data should be added only where there is a current, legitimate reason to use it.
Why JSON-LD is usually the practical choice
Google supports JSON-LD, Microdata and RDFa for structured data. For most modern websites, JSON-LD is the easiest of the three to manage because the markup can be maintained separately from the visible HTML structure.
That becomes especially useful when information is generated from templates or centralized data systems. Product details, article authors, business information and other fields can be mapped into JSON-LD rather than manually embedded throughout individual elements on every page.
JSON-LD does not receive special AI-search treatment simply because it is JSON-LD. It is a technical format. Its value comes from accurately representing information using an appropriate vocabulary.
If your implementation becomes more sophisticated and needs to connect organizations, people, products or other entities across a structured graph, see our guide to entity mapping and linked data for SEO.
How to choose which structured data to implement
A strong schema strategy starts with the page, not with a list of available Schema.org types.
- Identify what the page actually represents. Is it primarily a product, article, business location, person profile, event, video or something else?
- Check whether Google supports a relevant Search feature. If Google visibility is the objective, use current Search Central documentation rather than an old plugin checklist or a generic list from Schema.org.
- Confirm that the required information is visible and accurate. Do not use markup to describe content users cannot access or facts that differ from the page.
- Choose the implementation format. JSON-LD is generally the easiest option when your platform supports it cleanly.
- Validate before deployment. Test Google-supported features with the Rich Results Test and use Schema.org validation when checking broader vocabulary and syntax.
- Deploy carefully and monitor. Template changes, plugins, content updates and data integrations can all break otherwise valid markup later.
For large websites, the difficult part is rarely writing one JSON-LD block. It is building a system that keeps millions of generated values accurate as products, people, locations and page templates change. That enterprise-scale implementation deserves its own architecture rather than being reduced to a paragraph in a beginner guide.
Structured data is only as reliable as the data behind it
Valid syntax cannot rescue an incorrect fact.
A Product object with an old price is still wrong. An Organization object containing a former address is still wrong. An Article object attributing content to the wrong person is still wrong. A perfectly valid JSON-LD block can therefore create a data-quality problem even when every bracket and property passes validation.
This is why structured data should draw from dependable source systems wherever practical. Our guide to data consistency for SEO and AI search explains how to govern those underlying facts across websites, feeds, profiles and other systems.
Validate for the outcome you actually want
Different validation tools answer different questions.
Google's Rich Results Test checks whether markup can support Google Search rich-result features recognized by the tool. It can also preview some search appearances.
The Schema Markup Validator is broader. It can validate Schema.org syntax and vocabulary even when the markup does not correspond to a special Google Search feature.
A page can therefore contain valid Schema.org markup that does not produce a Google rich result. That is not necessarily an error. The important question is whether the markup has a legitimate purpose and accurately represents the page.
After deployment, monitor Search Console for structured-data issues associated with supported features. Also check rendered pages after CMS, plugin and template releases; structured data often breaks because the system generating it changed, not because an SEO intentionally edited the JSON-LD.
Common schema mistakes
Marking up information that is not visible or representative of the page. Google expects structured data to describe the page users actually encounter.
Assuming every Schema.org type produces a Google feature. Schema.org's vocabulary is broader than Google's rich-result support.
Using stale values. Prices, availability, dates, authors, locations and other dynamic information need the same maintenance discipline as visible content.
Adding schema merely because a plugin offers it. An available checkbox is not an SEO strategy. Choose markup based on the page and the supported use case.
Allowing multiple systems to generate conflicting markup. A page can legitimately contain several structured-data objects, but plugins, themes and custom code sometimes output competing versions of the same organization, product or article with different values.
Treating validation as proof of performance. Passing a validator proves that the markup meets particular technical requirements. It does not guarantee a rich result, ranking increase, Knowledge Panel or AI citation.
Using structured data to make claims the page does not support. Markup should describe reality, not embellish it.
How should you measure schema performance?
Measurement should match the reason the markup was implemented.
If the objective is eligibility for a Google rich-result feature, monitor the relevant Search Console enhancement and search-performance reporting where available. Look for valid-item coverage, errors, impressions, clicks and any meaningful change in search appearance.
If the implementation addresses a data-quality problem, measure whether the inaccurate or conflicting values are gone.
If schema was part of a larger page improvement, do not automatically assign every ranking, CTR or conversion change to the markup. Content changes, titles, rankings, SERP layout, seasonality and many other variables can move at the same time.
And if you are interested in AI visibility, measure AI visibility directly. Track whether the brand and pages appear for relevant prompts and whether AI systems cite or refer traffic to the site. Do not infer that success from the existence of JSON-LD.
Where schema fits in LSEO's technical architecture
Structured data works best as one part of a coherent information system.
Your website taxonomy determines how content and entities are classified. Data governance keeps the underlying facts accurate. Entity mapping defines meaningful relationships among people, organizations, products and other things. Structured data gives you a standardized way to publish appropriate parts of that information in machine-readable form.
None of those layers replaces the visible website.
A technically perfect Article object does not turn a weak article into a useful one. Product markup does not fix a bad product detail page. Organization schema does not create organizational authority by itself.
The most durable strategy is simple: build useful pages, maintain accurate facts, use clear architecture, and add structured data where it accurately describes the content or enables a documented feature.
What schema markup means in an AI-search world
The rise of AI-assisted discovery has made machine-readable information more interesting, but it has also produced a large amount of speculation.
For Google, the evidence is straightforward: traditional SEO fundamentals remain foundational to AI Overviews and AI Mode, and no special structured-data markup is required for inclusion.
For the broader AI ecosystem, marketers should resist converting a reasonable technical practice into an unsupported ranking theory. An AI product may process information from the web in ways that change over time, and there is no universal schema standard that forces an independent system to cite, quote or recommend a particular source.
Use schema because it solves a real information or search-presentation problem. Then measure the outcomes that matter to your business.
Need Help With Structured Data, Technical SEO or AI Search?
LSEO helps organizations connect technical SEO, structured data, site architecture and AI visibility measurement without relying on unsupported shortcuts.
Frequently Asked Questions
What is schema markup?
Schema markup is structured data that uses a standardized vocabulary, commonly Schema.org, to describe information on a webpage in machine-readable form. It can identify things such as products, organizations, articles, people, events and their properties.
Is schema markup the same as JSON-LD?
No. Schema.org provides vocabulary for describing things and relationships. JSON-LD is one format for publishing structured data using that vocabulary. Google also supports Microdata and RDFa, although JSON-LD is generally its recommended format when practical.
Does schema markup improve SEO rankings?
Schema should not be treated as a direct ranking boost. Its clearest Google Search benefit is helping Google understand structured information and making eligible pages available for supported rich-result features. Those features are not guaranteed to appear simply because the markup is valid.
Does schema markup help with Google AI Overviews or AI Mode?
Google says no special Schema.org markup is required for its generative AI search features. Structured data can still be useful as part of normal SEO and for supported rich results, but it should not be presented as a requirement or guaranteed optimization for AI Overviews or AI Mode.
What is the best structured-data format for SEO?
Google supports JSON-LD, Microdata and RDFa. JSON-LD is generally the most practical choice because Google recommends it when possible and it is often easier to maintain separately from the visible HTML.
Should every page have schema markup?
No. Add structured data when it accurately represents the page and has a legitimate purpose. Some pages have obvious use cases, such as products, articles, videos, business locations or profile pages. Adding unnecessary markup simply because a Schema.org type exists does not create additional SEO value.
Is FAQ schema still useful?
FAQPage remains a Schema.org type, but Google significantly restricted FAQ rich results. For most commercial websites, FAQPage markup should not be implemented with the expectation of receiving an expanded FAQ result in Google Search. Useful FAQ content can still be valuable to users regardless of whether FAQPage markup is present.
How do I test schema markup?
Use Google's Rich Results Test when you are implementing a structured-data feature supported by Google Search. For broader Schema.org syntax and vocabulary, use the Schema Markup Validator. After deployment, monitor the relevant Search Console reports and periodically verify that templates continue producing accurate markup.
Can schema markup increase AI citations?
There is no universal documented rule showing that schema markup directly increases citations across AI systems. Use structured data for its established purposes and measure AI citation visibility separately rather than assuming a causal relationship.