Schema.org Usage Statistics: What Google's First Adoption Dataset Reveals About Structured Data Strategy & Rankings

The SEO Industry Finally Has Real Structured Data Evidence

For many years, SEOs and web developers implemented structured data based on belief. They adapted to Google's guidelines, imitated their rivals, and hoped that their schema markup would be converted into rich results. Unfortunately, the feedback loop was almost non-existent. That is about to change.

On June 4, 2026, the Google Plus Schema.org community published the first-ever public dataset of structured data usage statistics about adoption at a crawl scale drawn straight from the Google index, covering millions of domains. It is the kind of absolute evidence that the SEO industry has never had the privilege to access before. And what it discloses about the actual usage of schema on the web should cause a lot of practitioners to completely rethink their strategy.

What Is the Schema.org Usage Statistics Dataset?

This dataset is a collaborative project by Google and the Schema.org community aimed at exposing how structured data vocabulary gets used on the web publicly. Ryan Levering, a Google software engineer, shared on LinkedIn that Dan Brickley, the main coordinator of Schema.org, had been requesting Google to release usage statistics for a long time. The project eventually got the green light from leadership and was released, and it is significant because no other search engine crawl can compete with Google's indexing depth.

This dataset has been designed with three attributes per record: term type (either a Type like Product or a Property like Price), the official Schema.org URI for that term, and a domain count bucket showing the range of adoption. Buckets start at less than 1,000 domains and go up to more than 10 million. Instead of publishing exact numbers that change constantly due to crawl noise, Schema.org uses these range values to give a fairly stable and meaningful indication of adoption while not allowing reverse engineering of crawling patterns.

One very important detail in the method: these counts are aggregated at the domain level, not per page or by the number of schema objects installed. A publisher with 500 pages that use Product schema will still be counted as one domain. This is a conscious choice; it measures the breadth of adoption rather than the depth of usage, making comparisons between different terms more trustworthy.

The release of May 2026 includes data for 958 item types and 4,587 predicates (properties), making 5,545 entries in total. The files can be downloaded in CSV, JSON, and a JSON summary format from the official Schema.org GitHub repository, and monthly updates are expected. Adoption data will also be shown on term pages at Schema.org for feature developers, accompanied by documentation and implementation guidelines.

The 12 Types That Built the Web's Schema Layer

First and foremost, this dataset demonstrates that the structured data web has an incredibly narrow base. Only 12 item types are in the top adoption bucket, over 10 million domains, and none of them are what most SEOs would call "content types." They are structural types:

BreadcrumbList, EntryPoint, ImageObject, ListItem, Organization, Person, PropertyValueSpecification, ReadAction, SearchAction, Thing, WebPage, WebSite.

That is the first revelation the data forces us to confront: the most popular schema on the internet is not Product, Article, or Recipe. It is the hidden scaffolding, the types that describe how a website is structured, not what it contains. WebPage, WebSite, Organization, and BreadcrumbList appear on over 10 million domains because they are the standard output of every major CMS and SEO plugin, applied globally to entire sites rather than individual pieces of content.

This is a major point: infrastructure has been the main force behind most of the web's structured data footprint rather than deliberate editorial strategy. SearchAction at 10M+ domains is not the outcome of SEOs meticulously implementing sitelinks search boxes. It is Yoast, RankMath, and other plugins adding it by default. The web's schema layer was constructed less by human decision-making than by a handful of widely adopted defaults.

When SEOs talk about "implementing schema," they are often working at a level above this invisible foundation, and that foundation hardly tells Google anything about what makes one site's content different from another's.

The CMS Plugin Problem: Why Most Schema Looks Identical

WordPress, Yoast, and Schema Monoculture

These 12 universal types indicate a fundamental issue in the way structured data has been developed. WordPress operates on approximately 43% of online content. Yoast SEO has more than 10 million active users. When a plugin creates schema automatically for each website that installs it, the adoption statistics of that plugin's default output appear very high. But it does not communicate anything new to Google.

Take Organization schema, used on over 10 million domains. At first glance, this seems like a massive implementation of a valuable data type. In reality, a large part of this is just the same boilerplate: identical fields, the same depth, coming from the same plugin template. When millions of websites produce Organization structures that are almost the same except for the name and URL, it becomes very difficult for Google to use this data to identify semantic meaning. The type is there. The actual differentiating content is missing.

Generic Schema Loses Value at Scale

Schema markup is most useful when it is specific and rare. A type or property that appears on most sites becomes background noise that search algorithms filter out rather than a source of signals. The more a type is used as a generic template, the less unique the information is within each deployment.

The data illustrate this point clearly. FAQPage is likely among the 1M-10M sites, a rather high adoption. Yet a large proportion of those implementations are plugin-generated, with templated questions and content that is duplicated or very thin. Google has pulled back on offering FAQ-rich snippets in quite a few contexts, partly perhaps because the quality of the average FAQPage content has deteriorated as the scale of automated rollout has increased. Mass adoption and quality did not go hand in hand.

The insight is clear: schema types that have reached monoculture status, deployed in exactly the same way by millions of sites, have largely lost their potential to differentiate. They are table stakes, not a source of competitive advantage. Continuing to optimise Organization or WebSite markup is like optimising a commodity.

The 76.9% Opportunity: Why Unused Schema Types Matter Most

Low Adoption Equals Low Competition

This is where the data shifts from diagnostic to genuinely strategic. Of the 958 item types in the May 2026 release, 485 are in the sub-1,000 domain group, meaning fewer than 1,000 domains worldwide are using them. That is 50.6% of all schema types having almost zero adoption. When the buckets of 1K-10K and 10K-100K are added, approximately 76.9% of all schema types are used on fewer than 100,000 domains.

This is not a problem of vocabulary. Schema.org is not defining types that are so obscure the web has no use for them. It is a matter of visibility and tooling: SEOs bet on the types that plugins implement, documentation focuses on, and rich results reward, which is really just a very small part of what is available. The rest of the vocabulary lies mostly unused, and for sites operating in those verticals, there is practically no schema-level competition.

Low adoption does not mean low value to Google. It signifies low noise, which renders each implementation more outstanding. When Google sees that Physician, LegalService, or FinancialProduct schema is implemented thoroughly and correctly, it is exactly because the adoption floor is so limited that the signal becomes noticeable. The signal-to-noise ratio is the complete opposite of that for ubiquitous types.

Industry-Specific Examples: Medical, Legal, Financial

Medical, legal, and financial markets are among the most lucrative in the entire dataset and the most critical from a search quality perspective, since these are areas where Google has always prioritised trust and authority signals.

Healthcare. Schema.org breaks down the MedicalEntity category into different types: Physician, MedicalCondition, Drug, Hospital, MedicalClinic, MedicalProcedure, and more. The majority of them have fewer than 10,000 domains using them. For a healthcare publisher or hospital system with actual medical content, deploying proper physician profiles with specialty details, board certifications, and hospital affiliations, or MedicalCondition pages with properly referenced symptoms and treatments, generates a schema presence that almost no competitor has. In a YMYL category where Google is actively seeking to judge expertise and trustworthiness, unique structured data is a powerful signal.

Law. LegalService, Attorney, and LegalCase are types that are not frequently used even though the law vertical is one of the highest CPL in search advertising and one of the most competitive in organic search. A law firm or legal publisher that deploys Attorney schema with accurate practice area specificity, bar admissions, and geographic coverage at scale is communicating in a structured language that very few organic competitors are using. This is particularly relevant for firms with multiple locations where local differentiation plays a role.

Finance. FinancialProduct, LoanOrCredit, BankAccount, and InvestmentFund are financial schema types that are incredibly rare even though the financial sector is both large and competitive. Most financial websites implement general Organization and WebPage markup only, while the entire financial vocabulary remains almost completely unimplemented. In a field where Google's quality raters are specifically instructed to evaluate financial authority, the combination of strong E-E-A-T signals and accurate financial schema markup remains an underutilised hybrid.

The pattern is consistent across all three sectors: these are high-value verticals where Google deeply cares about trust, expertise, and structured signals, yet the specialised schema vocabulary for each remains at near-zero adoption. That is an extraordinary strategic opening for any site willing to implement with precision.

What This Means for Your Schema Strategy

The dataset completely changes how we should talk about structured data. Instead of asking "Am I using the recommended types correctly?" the real question becomes "What can I implement that most of my competitors are not doing and that Google has very little structured information about?"

That means performing a schema audit by comparing your implementation against the adoption data. If your entire implementation falls within the 1M+ domain buckets, you are doing infrastructure, not strategy. The strategic move is discovering the Schema.org types that complement your content, particularly in specific, expert, or YMYL topics, and implementing them accurately and comprehensively where the competitive floor is virtually zero.

The dataset will be refreshed monthly. That means there is a real, clear, and quantifiable first-mover advantage opportunity in low-adoption types. For SEOs ready to move beyond the defaults, Google has just published a roadmap to the territory no one else is working.

At Oddtusk, we build schema strategies grounded in semantic precision, entity-first architecture, and validated structured data. Our SEO services and content marketing solutions integrate schema deployment with topical authority mapping, E-E-A-T optimisation, and AI readiness to future-proof your visibility. Ready to turn your content into a structured competitive advantage? Let's build your schema strategy together.