S SCHEMA.BIZ

State of Schema 2026: Structured Data on the Top 10,000 Sites

Last updated:

We crawled the home pages of the 10,000 most popular sites on the Tranco list in November 2026 and ran their markup through the Schema.biz validator. Of the 7,580 sites that answered, we could analyze 5,687. This page is what we found, with every chart free to cite and the per-site data free to download.

Key findings

  • 58.4% of the sites we could analyze publish schema.org markup on their home page or the one linked page we fetched. JSON-LD is the format on 55.4%, Microdata on 12.5% and RDFa on 0.8%.
  • 27.7% still ship markup for a Google feature that no longer exists. The biggest is the sitelinks search box, retired on November 21, 2024: 22.9% of sites still have its SearchAction. FAQ rich results ended on May 7, 2026, and 6.8% of sites still mark up an FAQPage.
  • The most common problem is a missing property Google recommends, on 70.8% of the sites with markup. The most common error, an item without @type, is on 2.9%.
  • 18.9% of robots.txt files block at least one AI crawler from the home page; CCBot is blocked most often (15.2%).
  • 17.4% of sites have an llms.txt, and 78.5% of those pass our validator.
  • 21.7% of the sites that answered didn't let us in: 15.7% refused the request and 4.1% disallow our crawler in robots.txt. We count them as not fetched, never as sites without markup.

Formats: JSON-LD leads, Microdata hangs on

Google reads all three formats and recommends JSON-LD where a site's setup allows it. On the top sites JSON-LD is the norm: 46% of analyzed sites use it alone, 9.4% mix it with Microdata or RDFa, and 3.9% use Microdata or RDFa without any JSON-LD. Converting the older markup is mechanical; the Microdata to JSON-LD Converter does it from a page's address, and the formats guide compares the three.

Structured data formats on the top sites % of the 5,687 sites we could analyze
  • Any schema.org markup 58.4%
  • JSON-LD 55.4%
  • Microdata 12.5%
  • RDFa 0.8%
  • JSON-LD with Microdata or RDFa 9.4%
Table: counts behind the chart
Structured data formats on the top sites Sites Share
Any schema.org markup 3,319 58.4%
JSON-LD 3,149 55.4%
Microdata 713 12.5%
RDFa 46 0.8%
JSON-LD with Microdata or RDFa 533 9.4%
Cite or embed

Free to use under CC BY 4.0, with a link to this chart as the attribution. Open the PNG.

<figure>
  <a href="https://schema.biz/learn/state-of-schema-2026/#formats"><img src="https://schema.biz/data/state-of-schema-2026/formats.png" alt="Structured data formats on the top sites (% of the 5,687 sites we could analyze): Any schema.org markup 58.4%, JSON-LD 55.4%, Microdata 12.5%, RDFa 0.8%, JSON-LD with Microdata or RDFa 9.4%." width="1200"></a>
  <figcaption>Structured data formats on the top sites. Source: <a href="https://schema.biz/learn/state-of-schema-2026/#formats">State of Schema 2026, Schema.biz</a> (CC BY 4.0).</figcaption>
</figure>

The types the top sites use

The types that describe the site lead: Organization and WebSite say who runs it and what it's called, ahead of page-level types. Shares count a site once if any page we fetched has the type, nested ones included (a PostalAddress inside an Organization counts). The type pages in our schema type reference show these figures for each type found on at least 50 sites.

The most used schema.org types % of the 5,687 sites we could analyze
  • Organization 40%
  • WebSite 36.4%
  • ImageObject 31.4%
  • ListItem 28.1%
  • WebPage 27.5%
  • BreadcrumbList 27.2%
  • SearchAction 24.3%
  • Person 15.8%
  • PostalAddress 13.6%
  • EntryPoint 13.4%
  • ContactPoint 12.7%
  • ReadAction 7.5%
  • PropertyValueSpecification 7.3%
  • Answer 6.9%
  • FAQPage 6.9%
Table: counts behind the chart
The most used schema.org types Sites Share
Organization 2,276 40%
WebSite 2,068 36.4%
ImageObject 1,788 31.4%
ListItem 1,596 28.1%
WebPage 1,566 27.5%
BreadcrumbList 1,545 27.2%
SearchAction 1,381 24.3%
Person 897 15.8%
PostalAddress 775 13.6%
EntryPoint 761 13.4%
ContactPoint 722 12.7%
ReadAction 429 7.5%
PropertyValueSpecification 414 7.3%
Answer 393 6.9%
FAQPage 393 6.9%
Cite or embed

Free to use under CC BY 4.0, with a link to this chart as the attribution. Open the PNG.

<figure>
  <a href="https://schema.biz/learn/state-of-schema-2026/#types"><img src="https://schema.biz/data/state-of-schema-2026/types.png" alt="The most used schema.org types (% of the 5,687 sites we could analyze): Organization 40%, WebSite 36.4%, ImageObject 31.4%, ListItem 28.1%, WebPage 27.5%, BreadcrumbList 27.2%, SearchAction 24.3%, Person 15.8%, PostalAddress 13.6%, EntryPoint 13.4%, ContactPoint 12.7%, ReadAction 7.5%, PropertyValueSpecification 7.3%, Answer 6.9%, FAQPage 6.9%." width="1200"></a>
  <figcaption>The most used schema.org types. Source: <a href="https://schema.biz/learn/state-of-schema-2026/#types">State of Schema 2026, Schema.biz</a> (CC BY 4.0).</figcaption>
</figure>

Markup for features Google retired

Google has retired several rich results since 2023, and markup written for them stays on pages long after. The sitelinks search box ended on November 21, 2024; FAQ rich results on May 7, 2026; How-to on September 13, 2023, and 0.4% of sites still have HowTo markup. Leftover markup does no harm: Google said in 2023 that structured data it doesn't use doesn't cause problems for Search. But it's weight in every page, it can still carry errors, and it's a sign nobody has looked at the markup in years. Our SearchAction and FAQPage pages say what each still does.

Markup for Google features that no longer exist % of the 5,687 sites we could analyze
  • Any retired feature 27.7%
  • Sitelinks search box · retired Nov 2024 22.9%
  • FAQ · retired May 2026 6.8%
  • How-to · retired Sep 2023 0.4%
  • Course info · retired Sep 2025 0.1%
  • Learning video · retired Sep 2025 0%
  • Claim review (fact check) · retired Sep 2025 0%
Table: counts behind the chart
Markup for Google features that no longer exist Sites Share
Any retired feature 1,576 27.7%
Sitelinks search box (retired Nov 2024) 1,305 22.9%
FAQ (retired May 2026) 388 6.8%
How-to (retired Sep 2023) 25 0.4%
Course info (retired Sep 2025) 3 0.1%
Learning video (retired Sep 2025) 1 0%
Claim review (fact check) (retired Sep 2025) 1 0%
Cite or embed

Free to use under CC BY 4.0, with a link to this chart as the attribution. Open the PNG.

<figure>
  <a href="https://schema.biz/learn/state-of-schema-2026/#retired"><img src="https://schema.biz/data/state-of-schema-2026/retired.png" alt="Markup for Google features that no longer exist (% of the 5,687 sites we could analyze): Any retired feature 27.7%, Sitelinks search box 22.9%, FAQ 6.8%, How-to 0.4%, Course info 0.1%, Learning video 0%, Claim review (fact check) 0%." width="1200"></a>
  <figcaption>Markup for Google features that no longer exist. Source: <a href="https://schema.biz/learn/state-of-schema-2026/#retired">State of Schema 2026, Schema.biz</a> (CC BY 4.0).</figcaption>
</figure>

Which Google features the markup qualifies for

A feature counts when the markup of the right type has every property Google requires for it, as the Schema Markup Validator checks it. Most of what qualifies is site-level: the Organization's logo and details, site names and breadcrumbs. Page-level features such as products and articles depend on the inner page we fetched, so they're undercounted (see the limitations).

Sites whose markup qualifies for a Google feature % of the 5,687 sites we could analyze
  • Organization 38.1%
  • Site name 31.1%
  • Breadcrumb 22.1%
  • Article 13.9%
  • Video 3.3%
  • Course list 2.8%
  • Carousel 2.7%
  • Product snippet 1.9%
  • Speakable 1.4%
  • Review snippet 1.3%
Table: counts behind the chart
Sites whose markup qualifies for a Google feature Sites Share
Organization 2,166 38.1%
Site name 1,771 31.1%
Breadcrumb 1,255 22.1%
Article 788 13.9%
Video 189 3.3%
Course list 159 2.8%
Carousel 154 2.7%
Product snippet 108 1.9%
Speakable 80 1.4%
Review snippet 74 1.3%
Cite or embed

Free to use under CC BY 4.0, with a link to this chart as the attribution. Open the PNG.

<figure>
  <a href="https://schema.biz/learn/state-of-schema-2026/#features"><img src="https://schema.biz/data/state-of-schema-2026/features.png" alt="Sites whose markup qualifies for a Google feature (% of the 5,687 sites we could analyze): Organization 38.1%, Site name 31.1%, Breadcrumb 22.1%, Article 13.9%, Video 3.3%, Course list 2.8%, Carousel 2.7%, Product snippet 1.9%, Speakable 1.4%, Review snippet 1.3%." width="1200"></a>
  <figcaption>Sites whose markup qualifies for a Google feature. Source: <a href="https://schema.biz/learn/state-of-schema-2026/#features">State of Schema 2026, Schema.biz</a> (CC BY 4.0).</figcaption>
</figure>

The most common problems

Shares here are of the 3,319 sites with any schema.org markup. Most problems are warnings: the markup parses and Google can use it, but a value is in the wrong form or a property Google reads is missing. Errors that stop the markup counting at all are rare.

The most common problems in the markup % of the 3,319 sites with schema.org markup
  • A missing property Google recommends · REC_MISSING 70.8%
  • A breadcrumb with fewer than two items · BC_TOO_FEW 19.7%
  • A URL that isn't absolute or valid · VALUE_URL 18.1%
  • A missing property Google requires · REQ_MISSING 15.3%
  • An image smaller than Google's minimum · IMAGE_TOO_SMALL 8.1%
  • A property used on the wrong type · PROP_NOT_ON_TYPE 6.4%
  • An app without a rating or review · APP_NO_RATING 5%
  • A property schema.org doesn't define · PROP_UNKNOWN 5%
  • A date that isn't ISO 8601 · VALUE_DATE 3.9%
  • A nested object without @type · TYPE_MISSING_NESTED 3.3%
Table: counts behind the chart
The most common problems in the markup Sites Share
A missing property Google recommends (REC_MISSING) 2,349 70.8%
A breadcrumb with fewer than two items (BC_TOO_FEW) 655 19.7%
A URL that isn't absolute or valid (VALUE_URL) 602 18.1%
A missing property Google requires (REQ_MISSING) 509 15.3%
An image smaller than Google's minimum (IMAGE_TOO_SMALL) 270 8.1%
A property used on the wrong type (PROP_NOT_ON_TYPE) 212 6.4%
An app without a rating or review (APP_NO_RATING) 167 5%
A property schema.org doesn't define (PROP_UNKNOWN) 165 5%
A date that isn't ISO 8601 (VALUE_DATE) 130 3.9%
A nested object without @type (TYPE_MISSING_NESTED) 110 3.3%
Cite or embed

Free to use under CC BY 4.0, with a link to this chart as the attribution. Open the PNG.

<figure>
  <a href="https://schema.biz/learn/state-of-schema-2026/#issues"><img src="https://schema.biz/data/state-of-schema-2026/issues.png" alt="The most common problems in the markup (% of the 3,319 sites with schema.org markup): A missing property Google recommends 70.8%, A breadcrumb with fewer than two items 19.7%, A URL that isn't absolute or valid 18.1%, A missing property Google requires 15.3%, An image smaller than Google's minimum 8.1%, A property used on the wrong type 6.4%, An app without a rating or review 5%, A property schema.org doesn't define 5%, A date that isn't ISO 8601 3.9%, A nested object without @type 3.3%." width="1200"></a>
  <figcaption>The most common problems in the markup. Source: <a href="https://schema.biz/learn/state-of-schema-2026/#issues">State of Schema 2026, Schema.biz</a> (CC BY 4.0).</figcaption>
</figure>

Organization markup: logos are common, sameAs less so

Google uses an Organization's logo and its sameAs links to social profiles and knowledge bases to understand who runs a site. Of the 2,596 sites with Organization markup, 88.4% give a logo and 69.1% give sameAs links; 62% have name, url, logo and sameAs together. The Organization schema page lists what Google reads.

What the Organization markup includes % of the 2,596 sites with Organization markup
  • url 89.6%
  • logo 88.4%
  • sameAs (social profiles) 69.1%
  • name, url, logo and sameAs 62%
Table: counts behind the chart
What the Organization markup includes Sites Share
url 2,325 89.6%
logo 2,295 88.4%
sameAs (social profiles) 1,794 69.1%
name, url, logo and sameAs 1,610 62%
Cite or embed

Free to use under CC BY 4.0, with a link to this chart as the attribution. Open the PNG.

<figure>
  <a href="https://schema.biz/learn/state-of-schema-2026/#organization"><img src="https://schema.biz/data/state-of-schema-2026/organization.png" alt="What the Organization markup includes (% of the 2,596 sites with Organization markup): url 89.6%, logo 88.4%, sameAs (social profiles) 69.1%, name, url, logo and sameAs 62%." width="1200"></a>
  <figcaption>What the Organization markup includes. Source: <a href="https://schema.biz/learn/state-of-schema-2026/#organization">State of Schema 2026, Schema.biz</a> (CC BY 4.0).</figcaption>
</figure>

AI crawlers in robots.txt

We read each site's robots.txt for the AI crawler tokens their vendors document and checked whether each may fetch the home page. 18.9% of the 7,475 files we could read block at least one. Blocking a training crawler such as GPTBot (14.5%) or Google-Extended (12.5%) doesn't stop the same company's search crawler; the robots.txt checker shows which is which. 367 files carry a Content-Signal line, a proposal RFC 9309 doesn't define.

AI crawlers blocked from the home page by robots.txt % of the 7,475 sites whose robots.txt we could read
  • Any AI crawler 18.9%
  • CCBot · Common Crawl 15.2%
  • GPTBot · OpenAI 14.5%
  • ClaudeBot · Anthropic 13.7%
  • meta-externalagent · Meta 12.6%
  • Google-Extended · Google 12.5%
  • Diffbot · Diffbot 12.2%
  • Amazonbot · Amazon 12.2%
  • Applebot-Extended · Apple 12.1%
  • PerplexityBot · Perplexity 8.6%
  • Google-CloudVertexBot · Google 8.4%
  • YouBot · You.com 8%
  • ChatGPT-User · OpenAI 7.6%
Table: counts behind the chart
AI crawlers blocked from the home page by robots.txt Sites Share
Any AI crawler 1,411 18.9%
CCBot (Common Crawl) 1,137 15.2%
GPTBot (OpenAI) 1,083 14.5%
ClaudeBot (Anthropic) 1,025 13.7%
meta-externalagent (Meta) 944 12.6%
Google-Extended (Google) 937 12.5%
Diffbot (Diffbot) 914 12.2%
Amazonbot (Amazon) 910 12.2%
Applebot-Extended (Apple) 903 12.1%
PerplexityBot (Perplexity) 640 8.6%
Google-CloudVertexBot (Google) 631 8.4%
YouBot (You.com) 601 8%
ChatGPT-User (OpenAI) 567 7.6%
Cite or embed

Free to use under CC BY 4.0, with a link to this chart as the attribution. Open the PNG.

<figure>
  <a href="https://schema.biz/learn/state-of-schema-2026/#ai-crawlers"><img src="https://schema.biz/data/state-of-schema-2026/ai-crawlers.png" alt="AI crawlers blocked from the home page by robots.txt (% of the 7,475 sites whose robots.txt we could read): Any AI crawler 18.9%, CCBot 15.2%, GPTBot 14.5%, ClaudeBot 13.7%, meta-externalagent 12.6%, Google-Extended 12.5%, Diffbot 12.2%, Amazonbot 12.2%, Applebot-Extended 12.1%, PerplexityBot 8.6%, Google-CloudVertexBot 8.4%, YouBot 8%, ChatGPT-User 7.6%." width="1200"></a>
  <figcaption>AI crawlers blocked from the home page by robots.txt. Source: <a href="https://schema.biz/learn/state-of-schema-2026/#ai-crawlers">State of Schema 2026, Schema.biz</a> (CC BY 4.0).</figcaption>
</figure>

llms.txt

llms.txt is a proposal, not a standard, and the major AI providers haven't said they read it. Still, 17.4% of the analyzed sites serve one at /llms.txt as text (we didn't count HTML pages served at that address). 78.5% of those pass the llms.txt validator. Our llms.txt guide explains the format.

llms.txt on the top sites % of the 5,687 sites we could analyze
  • Has an llms.txt 17.4%
  • Has one that passes our validator 13.7%
Table: counts behind the chart
llms.txt on the top sites Sites Share
Has an llms.txt 991 17.4%
Has one that passes our validator 778 13.7%
Cite or embed

Free to use under CC BY 4.0, with a link to this chart as the attribution. Open the PNG.

<figure>
  <a href="https://schema.biz/learn/state-of-schema-2026/#llms-txt"><img src="https://schema.biz/data/state-of-schema-2026/llms-txt.png" alt="llms.txt on the top sites (% of the 5,687 sites we could analyze): Has an llms.txt 17.4%, Has one that passes our validator 13.7%." width="1200"></a>
  <figcaption>llms.txt on the top sites. Source: <a href="https://schema.biz/learn/state-of-schema-2026/#llms-txt">State of Schema 2026, Schema.biz</a> (CC BY 4.0).</figcaption>
</figure>

Does rank make a difference?

The top of the Tranco list is full of infrastructure domains (CDNs, ad servers, APIs) with no website, so fewer sites near the top could be analyzed. Among those that could, the share with markup by rank band:

Sites with schema.org markup, by Tranco rank % of the analyzed sites in each rank band
  • Ranks 1–1,000 · 471 sites 59.9%
  • Ranks 1,001–5,000 · 2,150 sites 58.6%
  • Ranks 5,001–10,000 · 3,066 sites 58%
Table: counts behind the chart
Sites with schema.org markup, by Tranco rank Sites Share
Ranks 1–1,000 (471 sites) 282 59.9%
Ranks 1,001–5,000 (2,150 sites) 1,260 58.6%
Ranks 5,001–10,000 (3,066 sites) 1,777 58%
Cite or embed

Free to use under CC BY 4.0, with a link to this chart as the attribution. Open the PNG.

<figure>
  <a href="https://schema.biz/learn/state-of-schema-2026/#rank"><img src="https://schema.biz/data/state-of-schema-2026/rank.png" alt="Sites with schema.org markup, by Tranco rank (% of the analyzed sites in each rank band): Ranks 1–1,000 59.9%, Ranks 1,001–5,000 58.6%, Ranks 5,001–10,000 58%." width="1200"></a>
  <figcaption>Sites with schema.org markup, by Tranco rank. Source: <a href="https://schema.biz/learn/state-of-schema-2026/#rank">State of Schema 2026, Schema.biz</a> (CC BY 4.0).</figcaption>
</figure>

Methodology

  • Sample. The first 10,000 registrable domains of the Tranco list Q2VV4 (November 6, 2026), which combines Chrome UX Report, Cloudflare Radar, Cisco Umbrella, Majestic and Farsight rankings over 30 days. Domains whose home page ends on the same host as a higher-ranked one (country domains of one site) count once: 453 did.
  • What we fetched. From each site, robots.txt, the home page (https, then www), one page the home page links to (from its JSON-LD, else an article- or product-looking link) and /llms.txt. Crawled November 7, 2026 by SchemaBizBot from one machine, one request at a time per site, at least a second apart; how the crawler behaves.
  • robots.txt. Honored for SchemaBizBot on every host, redirects included, as RFC 9309 says: a missing file allows everything, and a server error or no answer disallows everything.
  • "Has markup". At least one schema.org type in JSON-LD, Microdata or RDFa, as the validator extracts it. RDFa in other vocabularies doesn't count.
  • Problems and features. The same engine as the Schema Markup Validator (schema.org 30.1), against Google's documented requirements, retired features included.
  • Denominators. Format, type, feature and llms.txt shares are of the 5,687 analyzed sites; problem shares of the 3,319 with markup; AI crawler shares of the 7,475 sites whose robots.txt we could read.
What happened when we asked for the home page % of the 7,580 sites that answered
  • Fetched and analyzed 75%
  • Refused (403, 429, a bot challenge) 15.7%
  • robots.txt disallows SchemaBizBot 4.1%
  • robots.txt unavailable 1.9%
  • Other error or not HTML 3.3%
Table: counts behind the chart
What happened when we asked for the home page Sites Share
Fetched and analyzed 5,687 75%
Refused (403, 429, a bot challenge) 1,193 15.7%
robots.txt disallows SchemaBizBot 310 4.1%
robots.txt unavailable 142 1.9%
Other error or not HTML 248 3.3%
Cite or embed

Free to use under CC BY 4.0, with a link to this chart as the attribution. Open the PNG.

<figure>
  <a href="https://schema.biz/learn/state-of-schema-2026/#access"><img src="https://schema.biz/data/state-of-schema-2026/access.png" alt="What happened when we asked for the home page (% of the 7,580 sites that answered): Fetched and analyzed 75%, Refused (403, 429, a bot challenge) 15.7%, robots.txt disallows SchemaBizBot 4.1%, robots.txt unavailable 1.9%, Other error or not HTML 3.3%." width="1200"></a>
  <figcaption>What happened when we asked for the home page. Source: <a href="https://schema.biz/learn/state-of-schema-2026/#access">State of Schema 2026, Schema.biz</a> (CC BY 4.0).</figcaption>
</figure>

Limitations

  • Two pages per site at most. A shop's product markup or a publisher's article markup is often on inner pages we didn't fetch, so page-level types are undercounted.
  • Raw HTML only: we run no JavaScript, so markup added by scripts is missed. 3.6% of home pages looked script-rendered (little text and an app root).
  • Sites that refused us are missing from every share, and we can't tell whether they mark up differently from the sites that let us in.
  • A rich result is never guaranteed: qualifying markup is what Google requires, not what it shows.

Download the data

state-of-schema-2026.csv: one row per site, 10,000 rows, licensed CC BY 4.0 (credit "State of Schema 2026, Schema.biz" with a link to this page). It holds counts and codes only, no page content. Columns:

ColumnWhat it holds
rankThe domain's rank on the Tranco list.
domainThe registrable domain, as the list gives it.
outcomeok (analyzed), refused, robots_disallowed, robots_unavailable, http_error, not_html, unreachable, blocked_address or duplicate.
detailFor other outcomes: the status code, the error class, or for a duplicate the domain it duplicates.
has_markup1 when a fetched page has schema.org markup.
json_ld_blocksJSON-LD blocks on the fetched pages.
microdata_itemsTop-level Microdata items.
rdfa_itemsTop-level schema.org RDFa items.
typesThe schema.org types found, nested ones included, space-separated.
errorsValidator error codes found, space-separated.
warningsValidator warning codes found, space-separated.
eligible_featuresGoogle features the markup qualifies for (Schema.biz feature ids).
retired_featuresRetired Google features the markup is still written for.
org_logo1 when the Organization markup has a logo (empty without Organization markup).
org_same_as1 when it has sameAs links.
robots_txtparsed, allow-all (no file) or disallow-all (server error or no answer).
ai_crawlers_blockedAI crawler tokens the robots.txt keeps off the home page.
llms_txtvalid, invalid, none or disallowed (robots.txt kept us from fetching it).
pages_checkedPages analyzed: 0, 1 (the home page) or 2 (and one linked page).

Site owners: to have your row removed, see the crawler page.

Check your own site

Run any page through the Schema Markup Validator for the same checks this study used, or the Schema Audit for a full report with fixes. The JSON-LD Generator writes new markup from a form, and the AI Readiness Checker tests robots.txt, snippets and llms.txt together.

How to cite this study

Schema.biz, "State of Schema 2026: Structured Data on the Top 10,000 Sites", November 7, 2026, https://schema.biz/learn/state-of-schema-2026/. Each chart has its own link and a PNG under "Cite or embed".