A data roundup of how much of the web carries schema markup, which formats and types sites use, what Google still supports, and what structured data does and doesn’t do for AI search. Every figure comes from the organisation that measured or published it: Web Data Commons (University of Mannheim) on the Common Crawl, HTTP Archive’s Web Almanac, Schema.org, Google Search Central and the WordPress.org plugin directory.
About half the web now carries schema markup: Web Data Commons found structured data on 51.25% of the 2.4 billion pages in the October 2024 Common Crawl, and the 2025 Web Almanac puts it on 50% of home pages, with JSON-LD the dominant format (Web Data Commons, Web Almanac 2025). The AI angle is where the myths are: Google says there is no special schema.org markup needed to appear in AI Overviews or AI Mode, while Microsoft has confirmed Bing uses schema.org markup to help Copilot tell articles, products, reviews and FAQs apart. Use this page if you need adoption benchmarks for a schema project, want to know which types are worth the effort after Google’s 2023 and 2025 cuts, or have to explain to a client what schema can and can’t do for AI answers. Skip it if you want crawl, robots.txt and Core Web Vitals data; those are on our technical SEO statistics page. If you want the markup written for you, Mentionova’s schema markup generator reads a page, finds the questions buyers ask AI engines about it and writes Article, BreadcrumbList and FAQPage JSON-LD, from $99/mo billed annually; our best schema markup generators ranking compares it with Schema App, WordLift, Yoast, Rank Math and the free options. Every figure below has a verbatim quote in our ledger.
Key takeaways
- 51.25% of the 2.4 billion pages in the October 2024 Common Crawl contained structured data (Web Data Commons)
- 50% of home pages carried structured data in 2025, up from 48% (desktop) and 49% (mobile) in 2024 (Web Almanac 2025)
- 41% of pages used JSON-LD in 2024, up from 34% in 2022 (Web Almanac 2024)
- 308K pay-level domains used FAQPage in 2024, against 19 in 2018 (Web Data Commons)
- 82% higher click-through rate for Nestlé pages shown as rich results, in a case study Google publishes (Google Search Central)
- 2.13% of desktop sites had a valid llms.txt file in 2025, the newer “AI-readable” file next to schema (Web Almanac 2025)
How much of the web uses structured data?
Two datasets measure this at web scale. Web Data Commons extracts every RDFa, Microdata, JSON-LD and Microformats statement from the Common Crawl; the Web Almanac uses HTTP Archive’s crawl of home pages and one inner page per site. Both land at about half.
- Web Data Commons found structured data on 1.3 billion of the 2.4 billion HTML pages in the October 2024 Common Crawl, or 51.25% (Web Data Commons).
- Those pages came from 16.5 million of the 37.5 million pay-level domains in the crawl, or 44.12% (Web Data Commons).
- The crawl parsed 2,391,039,772 HTML URLs, of which 1,245,622,627 had structured data triples (Web Data Commons).
- The extracted data sets hold 74 billion RDF quads (Web Data Commons).
- Web Data Commons counted 15,647,463,083 typed entities in the October 2024 extraction (Web Data Commons).
- 50% of home pages carried structured data on desktop and mobile in 2025, up from 48% and 49% in 2024 (Web Almanac 2025).
- Only 2% of crawls had structured data added via JavaScript; the rest ships it in the raw HTML (Web Almanac 2025).
- Web Data Commons has released 12 data sets from the Common Crawls of 2010 to 2023, plus the 2024 release, so adoption can be tracked over a decade (Web Data Commons).
| Measure (October 2024 Common Crawl) | Value |
|---|---|
| HTML pages parsed | 2,391,039,772 |
| Pages with structured data | 1,245,622,627 |
| Share of pages | 51.25% |
| Domains in crawl | 37,447,141 |
| Domains with structured data | 16,525,070 |
| Share of domains | 44.12% |
Source: Web Data Commons extraction report, October 2024 corpus. Domains are pay-level domains; subdomains are not counted separately.
Which structured data formats do sites use?
JSON-LD is the format Google recommends and the one growing fastest. RDFa still shows up on most pages because Open Graph tags are written in it, which inflates its share without making it the format SEO teams choose.
- Embedded JSON-LD was found on 11,562,359 domains and 833,818,654 URLs in the October 2024 crawl (Web Data Commons).
- Microdata was found on 7,599,792 domains and 574,648,578 URLs (Web Data Commons).
- RDFa was found on 474,635 domains and the hCard microformat on 3,522,517 (Web Data Commons).
- A JSON-LD page averages 57.54 triples (Web Data Commons).
- JSON-LD was on 41% of pages in 2024, up from 34% in 2022 (Web Almanac 2024).
- RDFa appeared on 66% of pages and Open Graph on 64%, mostly the same social sharing tags (Web Almanac 2024).
- Microdata appeared on 26% of pages and Twitter meta tags on 45% (Web Almanac 2024).
- Facebook meta tags fell to 7% of pages; Dublin Core and Microformats are each below 1% (Web Almanac 2024).
- JSON-LD was on 43% of desktop and mobile home pages in 2025, against Microdata on 17% of desktop and 16% of mobile home pages (Web Almanac 2025).
- On inner pages, JSON-LD reached 39% on desktop and 37% on mobile (Web Almanac 2025).
- Google recommends JSON-LD in most cases but says all 3 formats are equally fine when the markup is valid (Google Search Central).
Source: Web Almanac 2024, Structured Data chapter. A page can use several formats, so shares add up to more than 100%.
| Format | Domains | URLs |
|---|---|---|
| JSON-LD | 11,562,359 | 833,818,654 |
| Microdata | 7,599,792 | 574,648,578 |
| hCard | 3,522,517 | 179,698,044 |
| RDFa | 474,635 | 49,636,704 |
Source: Web Data Commons, October 2024 corpus, results per format.
Which schema types are most common?
The most common types describe the site itself (WebSite, Organization, BreadcrumbList), not the content. Content types such as Product, Article and FAQPage sit on a small share of pages, which is where the room to stand out is.
- schema.org/WebSite is the most common JSON-LD class, on 8,461,613 domains (Web Data Commons).
- Organization is on 6,574,284 domains and BreadcrumbList on 6,206,556 (Web Data Commons).
- Person is on 4,726,830 domains, the entity type behind author and expert markup (Web Data Commons).
- Article is on 2,358,095 domains and Product on 2,349,844 (Web Data Commons).
- LocalBusiness is on 1,263,114 domains and BlogPosting on 968,958 (Web Data Commons).
- WebSite appears on 37% of desktop home pages and 36% of mobile home pages in 2025 (Web Almanac 2025).
- SearchAction appears on 28% of home pages, even though Google retired the sitelinks search box it powered in November 2024 (Web Almanac 2025).
- Organization appears on 26.74% of desktop home pages and overtook WebPage for the first time in 2025 (Web Almanac 2025).
- WebSite on mobile home pages grew from 30% in 2022 to 35% in 2024 and 36% in 2025 (Web Almanac 2025).
- On inner pages, ListItem leads at 29% on desktop, and BreadcrumbList is on 28% of desktop inner pages (Web Almanac 2025).
- In the 2024 page-level JSON-LD data, WebSite was on 12.73% of mobile pages, Organization on 7.16% and LocalBusiness on 3.97% (Web Almanac 2024).
- BreadcrumbList JSON-LD was on 5.66% of pages and ItemList on 2.44% (Web Almanac 2024).
- BlogPosting was on 1.40% of pages, WebPage on 1.49% and Product on 0.77% (Web Almanac 2024).
- Industry types: AutoDealer on 1.09% of pages, Restaurant on 0.19% and Store on 0.17% (Web Almanac 2024).
- In Microdata, WebPage is on 8.34% of mobile pages and SiteNavigationElement on 6.42%, mostly theme-generated markup (Web Almanac 2024).
- Microdata Article is on 1.77% of pages and Product on 1.50% (Web Almanac 2024).
| JSON-LD class | Domains (Oct 2024) |
|---|---|
| WebSite | 8,461,613 |
| Organization | 6,574,284 |
| BreadcrumbList | 6,206,556 |
| Person | 4,726,830 |
| Article | 2,358,095 |
| Product | 2,349,844 |
| LocalBusiness | 1,263,114 |
| BlogPosting | 968,958 |
Source: Web Data Commons, October 2024 corpus, top JSON-LD classes by domain count.
| Type (2025) | Desktop home | Desktop inner |
|---|---|---|
| WebSite | 37% | 26% |
| SearchAction | 28% | 17% |
| Organization | 26.74% | 26% |
| BreadcrumbList | 21% | 28% |
| ListItem | 21% | 29% |
Source: Web Almanac 2025, SEO chapter, most popular structured data types on desktop home and inner pages.
Did FAQPage and HowTo survive Google’s cuts?
Google cut FAQ rich results to a small set of sites in August 2023 and ended HowTo rich results. Sites kept the markup anyway, and FAQPage kept growing, partly because other engines and AI systems still read it.
- Since August 8, 2023, FAQ rich results show only for well-known, authoritative government and health websites (Google Search Central).
- On September 14, 2023 Google extended the HowTo change to desktop, ending HowTo rich results on every device (Google Search Central).
- FAQPage was used by 19 pay-level domains in 2018 and by 308K in 2024 (Web Data Commons).
- FAQPage on desktop pages rose from 0.2% in 2022 to 0.6% in 2024, after the rich result cut (Web Almanac 2024).
- VideoObject, FAQPage and Event each appear on about 0.34% of pages in the 2024 JSON-LD data (Web Almanac 2024).
- HowTo schema adoption stayed below 1% on desktop and mobile (Web Almanac 2024).
Which structured data features has Google retired?
Google keeps trimming rich result types it says few sites use. The markup still works for other consumers; only the Google Search display goes away.
- On June 12, 2025 Google began phasing out 7 structured data features: Book Actions, Course Info, Claim Review, Estimated Salary, Learning Video, Special Announcement and Vehicle Listing (Google Search Central).
- On September 8, 2025 Search Console dropped reporting for 6 of those types (Google Search Central).
- Google does not guarantee that structured data will show in results, even when it passes the Rich Results Test (Google Search Central).
- A structured data manual action removes rich result eligibility but doesn’t affect how the page ranks in Google web search (Google Search Central).
| Date | Google change |
|---|---|
| Aug 8, 2023 | FAQ rich results limited to authoritative government and health sites |
| Sep 14, 2023 | HowTo rich results ended on desktop too |
| Nov 2024 | Sitelinks search box removed |
| Jun 12, 2025 | 7 features phased out, from Book Actions to Vehicle Listing |
| Sep 8, 2025 | Search Console reporting removed for 6 retired types |
Sources: Google Search Central, HowTo and FAQ changes, Simplifying the search results page; sitelinks search box date from the Web Almanac 2025.
Does schema markup help with AI search?
The honest answer is “it helps machines read the page, it is not a ticket into AI answers.” Google says nothing special is needed for its AI features; Microsoft says Bing’s models use schema.org to understand content types.
- Google: there are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary (Google Search Central).
- Google: there is no special schema.org structured data you need to add for AI features (Google Search Central).
- To be a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to show with a snippet (Google Search Central).
- Google says clicks from results pages with AI Overviews are higher quality, meaning users spend more time on the site (Google Search Central).
- Microsoft has confirmed Bing uses schema.org markup to help its models, including Bing Chat and Copilot, tell expert articles, products, reviews and FAQs apart (Web Almanac 2025).
- The Web Almanac’s structured data chapter names AI search engines such as Bing Chat and Google AI Overview as consumers of structured data (Web Almanac 2024).
What it means. Schema is cheap insurance, not a GEO strategy on its own. It makes your entities, prices and answers unambiguous for any system that reads the page, and Bing says it uses it. Whether ChatGPT, Perplexity or Google’s AI cites you still depends on the content and the sources those engines trust, which is why we track citations with an AI visibility tool rather than assume schema moved them. Our GEO statistics and Google AI Overviews statistics cover what does correlate with being cited.
Does structured data lift click-through rates?
Google’s own documentation publishes three case studies. They are vendor-reported, single-site results, so read them as “what is possible,” not an average.
- Rotten Tomatoes added structured data to 100,000 pages and measured a 25% higher click-through rate on enhanced pages (Google Search Central).
- The Food Network converted 80% of its pages to enable search features and saw a 35% increase in visits (Google Search Central).
- Nestlé measured an 82% higher click-through rate on pages shown as rich results than on non-rich result pages (Google Search Central).
| Case study | Change | Result |
|---|---|---|
| Rotten Tomatoes | 100,000 pages marked up | +25% CTR |
| Food Network | 80% of pages enabled | +35% visits |
| Nestlé | Rich result pages | +82% CTR |
Source: Google Search Central, Introduction to structured data, case studies published by Google.
How do sites link their entities with sameAs?
sameAs tells machines that your Organization or Person is the same entity as a profile elsewhere. It is one of the most direct ways to help a knowledge graph, and an AI engine, connect your brand to the right facts.
- sameAs links point to Facebook on 4.53% of pages and Instagram on 3.67% (Web Almanac 2024).
- LinkedIn appears in sameAs on 1.11% of pages, the main link for people and executives (Web Almanac 2024).
- Only 0.17% of pages link to Wikidata and 0.13% to Wikipedia, the sources that disambiguate entities best (Web Almanac 2024).
- schema.org is the JSON-LD @context in 20,960,693 implementations, against 11,973 for the next context (Web Almanac 2024).
- Organization grew from 1.3M to 2.2M pay-level domains between 2018 and 2019 in Web Data Commons’ long-run series (Web Data Commons).
- og:image is on 48% of desktop pages, the most-used Open Graph property (Web Almanac 2024).
How many sites publish llms.txt and AI crawler rules?
llms.txt is a proposed plain-text map of a site for language models. It is often discussed next to schema as “AI-readable” markup, so here is how far it has spread.
- 2.13% of desktop sites and 2.10% of mobile sites had a valid llms.txt in 2025 (Web Almanac 2025).
- 324,184 valid llms.txt files were found across the mobile sites analyzed (Web Almanac 2025).
- 39.6% of llms.txt files came from All in One SEO and 3.6% from Yoast SEO, so most are plugin-generated (Web Almanac 2025).
- gptbot is named in robots.txt on 20.9% of top-1k sites, 6.1% of top-100k sites and 4% of top-10M sites (Web Almanac 2025, Generative AI).
- 94.1% of about 12.9 million sites have a robots.txt with at least one directive (Web Almanac 2025, Generative AI).
For crawler-level data (which AI bots crawl most and who blocks them), see our AI crawler statistics. For the open crawl that many AI models train on, see our Common Crawl statistics.
How big is the schema.org vocabulary?
- schema.org has 826 types, 1540 properties, 19 datatypes, 96 enumerations and 544 enumeration members (Schema.org).
- Version 30.1, released 2026-09-16, added vocabulary for EU Digital Product Passports and common ecommerce product data (Schema.org).
- 30.1 added the Product properties consumerNotice, isOftenBoughtWith and specification for retail feed data (Schema.org).
- Version 30.0 shipped on 2026-03-19, and 29.4 (2025-12-08) added the ConferenceEvent type (Schema.org).
How do most sites add schema?
On WordPress, most schema comes from an SEO plugin rather than hand-written JSON-LD. Install counts from the WordPress.org directory show where that markup is generated.
- Yoast SEO has 10+ million active installations (WordPress.org).
- Rank Math has 4+ million active installations and lists 20+ pre-defined schema types (WordPress.org).
- All in One SEO has 2+ million active installations (WordPress.org).
- SEOPress has 300,000+ active installations (WordPress.org).
- Schema & Structured Data for WP & AMP has 100,000+ active installations and lists more than 35 schema types (WordPress.org).
- WordLift’s WordPress plugin has 300+ active installations (WordPress.org).
| WordPress plugin | Active installations |
|---|---|
| Yoast SEO | 10+ million |
| Rank Math | 4+ million |
| All in One SEO | 2+ million |
| SEOPress | 300,000+ |
| Schema & Structured Data for WP | 100,000+ |
| WordLift | 300+ |
Source: WordPress.org plugin directory pages for Yoast SEO, Rank Math, All in One SEO, SEOPress, Schema & Structured Data for WP and WordLift.
What to do with this data
- Cover the basics first. Organization, WebSite and BreadcrumbList are on millions of domains; missing them is the gap, not having them is the edge. Add sameAs links to LinkedIn, Wikidata and Wikipedia where they exist, since under 1% of pages do.
- Mark up what buyers ask. Product, Article and FAQPage sit on a small share of pages. FAQ rich results are gone for most sites, but Bing says its models use the markup, and the questions you mark up are the ones AI answers quote. Mentionova drafts FAQPage JSON-LD from the questions buyers ask ChatGPT, Perplexity, Claude and Gemini, then tracks whether those engines start citing the page.
- Ship JSON-LD in the HTML. Only 2% of sites inject schema with JavaScript; Google can read it, but other parsers may not.
- Don’t sell schema as a GEO shortcut. Google says no special markup is needed for AI Overviews or AI Mode. Measure citations before and after with a tracker from our AI citation tracking tools ranking.
- Keep markup matched to visible content. Google’s policies say not to mark up content readers can’t see, and a manual action removes rich result eligibility.
For the wider picture, see our technical SEO statistics, SEO statistics, AI search statistics and zero-click search statistics. To pick a generator, start with the best schema markup generators.
Definitions
- Structured data / schema markup: machine-readable statements about a page (its type, author, product, price, questions and answers), usually written with the schema.org vocabulary.
- JSON-LD: a JavaScript-notation format for structured data inside a script tag; Google’s recommended format.
- Microdata and RDFa: formats that add structured data as attributes on HTML elements.
- Rich result: a search result with extra display (stars, prices, FAQs) that structured data makes a page eligible for.
- Pay-level domain (PLD): a registrable domain such as example.com; Web Data Commons counts subdomains under it.
- sameAs: a schema.org property that links an entity to the same entity’s profiles elsewhere.
- llms.txt: a proposed plain-text file that summarises a site’s content for language models.
Frequently asked questions
What percentage of websites use schema markup?
About half. Web Data Commons found structured data on 51.25% of pages and 44.12% of domains in the October 2024 Common Crawl, and the Web Almanac found it on 50% of home pages in 2025.
What is the most common schema type?
WebSite. It is on 8,461,613 domains in JSON-LD and on 37% of desktop home pages, followed by Organization and BreadcrumbList.
What share of sites use JSON-LD?
41% of pages in the 2024 Web Almanac (up from 34% in 2022) and 43% of home pages in 2025. Web Data Commons found JSON-LD on 11,562,359 domains.
Does schema markup help with AI Overviews?
Not directly. Google says there is no special schema.org structured data needed to appear in AI Overviews or AI Mode; a page must be indexed and snippet-eligible.
Does Bing use schema for Copilot?
Yes, per the Web Almanac: Microsoft has confirmed Bing uses schema.org markup to help its models, including Copilot, tell articles, products, reviews and FAQs apart.
Is FAQ schema still worth adding?
For Google rich results, only government and health sites still get them. Adoption kept growing anyway, to 308K domains in 2024, because other systems read FAQ markup and the questions map to what buyers ask AI engines.
How much does structured data increase CTR?
Google’s published case studies show +25% CTR for Rotten Tomatoes, +35% visits for the Food Network and +82% CTR for Nestlé rich results. They are single-site results, not averages.
How many schema.org types are there?
826 types and 1540 properties as of version 30.1, released on September 16, 2026.
Which structured data types did Google retire?
FAQ (limited) and HowTo rich results in 2023, the sitelinks search box in November 2024, and seven more in June 2025: Book Actions, Course Info, Claim Review, Estimated Salary, Learning Video, Special Announcement and Vehicle Listing.
How many websites have llms.txt?
About 2%: 2.13% of desktop and 2.10% of mobile sites in the 2025 Web Almanac, with 39.6% of files generated by All in One SEO.
How we compiled these statistics
We fetched 18 primary source pages from Web Data Commons, HTTP Archive’s Web Almanac (2024 Structured Data, 2025 SEO and 2025 Generative AI chapters), Schema.org, Google Search Central and the WordPress.org plugin directory on November 9, 2026 and recorded 90 data points in docs/content/stats-ledgers/schema-markup-statistics.json. Every numbered figure above maps to a ledger row with a verbatim quote from the page we fetched. Web Data Commons and the Web Almanac measure different crawls (the Common Crawl’s 2.4 billion pages against HTTP Archive’s home and inner pages), so their percentages differ and we label each with its source rather than averaging them. Google’s CTR figures are case studies Google publishes, not a study we ran. We left out market-report sites, SEO vendor surveys and figures we couldn’t trace to the original publisher.
Sources
- Web Data Commons: extraction report, October 2024 Common Crawl corpus
- Web Data Commons: RDFa, Microdata, JSON-LD and Microformats data sets (trends 2012 to 2024)
- HTTP Archive: Web Almanac 2024, Structured Data chapter
- HTTP Archive: Web Almanac 2025, SEO chapter
- HTTP Archive: Web Almanac 2025, Generative AI chapter
- Schema.org: Organization of schemas
- Schema.org: Release history
- Google Search Central: AI features and your website
- Google Search Central: Introduction to structured data markup
- Google Search Central: Changes to HowTo and FAQ rich results (August 2023)
- Google Search Central: Simplifying the search results page (June 2025)
- Google Search Central: General structured data guidelines
- WordPress.org: Yoast SEO
- WordPress.org: Rank Math SEO
- WordPress.org: All in One SEO
- WordPress.org: SEOPress
- WordPress.org: Schema & Structured Data for WP & AMP
- WordPress.org: WordLift