Structured data in 2026-2027: what Google still reads, what it dropped, and JSON-LD templates for four site types

For most of a decade structured data was a way to decorate a search snippet. Google has since switched off most of those decorations, and the remaining value sits elsewhere: entity identification and fact grounding for AI answers. This note lists which Schema.org types Google still renders, which it ignores, and gives JSON-LD templates for four common site types.

The evolutionary shift: why structured data is needed in 2026–2027

From 2015 to 2022, structured data was seen by SEO specialists as a tool for visual snippet manipulation:

• marking up FAQPage to occupy half a smartphone screen;
• generating a 5-star rating via AggregateRating to increase CTR;
• inserting pseudo-prices or promotional badges.

Between 2023 and 2026, a fundamental shift occurred, driven by two processes:

1. Google restricted snippet extensions. Google sequentially disabled rich results for most manipulative schema types, including FAQ and HowTo.
2. The rapid growth of Generative Engine Optimization (GEO) and AI search. The emergence of Google AI Overviews, Perplexity, and ChatGPT Search shifted the focus of search from ten blue links to synthesizing a direct answer.

In modern systems, structured data serves two key functions:

                       ┌─────────────────────────────────────────────────────────┐
                       │             MODERN ROLE OF STRUCTURED DATA              │
                       └────────────────────────────┬────────────────────────────┘
                                                    │
                 ┌──────────────────────────────────┴──────────────────────────────────┐
                 ▼                                                                     ▼
   ┌───────────────────────────┐                                         ┌───────────────────────────┐
   │    VISUAL LAYER (SERP)    │                                         │      SEMANTIC LAYER       │
   ├───────────────────────────┤                                         ├───────────────────────────┤
   │ • Breadcrumbs (URL)       │                                         │ • Knowledge Graph binding │
   │ • Product cards           │                                         │ • E-E-A-T trust signals   │
   │   (prices, stock, returns)│                                         │   (authors, brands)       │
   │ • Video snippets          │                                         │ • Fact grounding for AI   │
   │ • Recipes and events      │                                         │                           │
   └───────────────────────────┘                                         └───────────────────────────┘

Structured data has moved from marketing to infrastructure: it is what lets crawlers and language models identify the entity on the page, keep its facts apart from a neighbour’s, and cite the page as a source.

Google specifics: clean search results and entity-based search

Google relies on the Schema.org standard in JSON-LD format. Its approach to displaying Rich Results has become significantly stricter.

The snippet purge

• FAQPage: Since August 2023, Google restricted FAQ accordions to trusted government (.gov) and authoritative medical sites. By 2026–2027, FAQPage markup completely stopped yielding visual snippets for commercial sites.
• HowTo: Completely disabled in search results, first on mobile, then desktop.
• Batch deprecation of 7 types (2025): Google officially ended snippet support for Course Info, Claim Review, Special Announcement, Book Actions, Estimated Salary, Learning Video, and Vehicle Listing.
• Stricter rules for AggregateRating: Self-serving reviews are banned. If a service site marks up a rating for itself on the homepage, Google does not display stars. Systematic abuse leads to manual actions for spammy structured markup.

What works in Google in 2026–2027

1. Product + Offer (Merchant Listings): The fastest developing markup type. Google shows prices, discount ranges, availability status, and new blocks: shippingDetails (shipping cost and speed) and hasMerchantReturnPolicy (return conditions). This provides free placement in the Google Shopping tab.
2. BreadcrumbList: A stable type. Replaces long technical URLs with a neat navigation path.
3. VideoObject: Mandatory for any video content. Secures placement in the video results block with Key Moments, duration, and a thumbnail.
4. ProfilePage and Person: Markup for experts and authors, introduced by Google to enhance content transparency under the E-E-A-T framework.
5. Organization with extended attributes: Google supports an extended list of organization properties: sameAs (links to Wikidata, registries, social networks), legalName, taxID, founder, parentOrganization for precise legal entity identification in the Knowledge Graph.

Impact on Google AI Overviews

Google has stated: “There is no special markup to get into AI Overviews.” However, the architecture of search LLMs relies on structured data as a low-cost method for fact validation. When a model maps statements on a page to a JSON-LD schema (price, author, datePublished), the page is more likely to be picked as a citation. That is a reading of how these systems work, not a published Google rule.

Direct versus indirect ranking effect

As a direct ranking factor

No. In Google’s algorithms, structured data alone does not carry isolated document weight. The presence of Schema.org does not automatically move a page from position 50 to 1.

Indirect influence mechanisms

1. CTR growth: A navigation path in BreadcrumbList, price, and availability status attract more attention than a plain text snippet.
2. Crawl budget economy: On large sites, search bots spend fewer resources parsing pages, accelerating the indexing of new materials.
3. Resistance to AI hallucinations: In the zero-click era, having structured facts in JSON-LD increases the chances of being selected as an authoritative citation source.

Summary matrix: what actually works

Element / Schema	Google today	What that means
BreadcrumbList	Works	Replaces raw URLs with a readable path.
Product + Offer	Works (price, stock, returns in the snippet)	Required for Merchant Listings.
AggregateRating (stars)	Works only in E-com Product; blocked for B2B/Local	Google banned self-serving reviews for services.
FAQPage	DISABLED (since late 2023)	Google removed FAQ accordions from all commercial sites. Present in code, absent in snippets.
LocalBusiness / Address	Indirect geo-signal	Google Local 3-Pack relies almost entirely on Google Business Profile, not just Schema.org.
Article / TechArticle	Works (date, Google Discover)	An image field ≥ 1200px wide is mandatory for Discover. dateModified provides a freshness signal.
VideoObject	Works	Requires thumbnailUrl, uploadDate, duration in ISO 8601 (e.g. PT4M30S).

Architectural templates for 4 site models

Model 1: B2B company or IT service

(Complex sales, expert services, long deal cycle)

Structured data map

• Homepage: Organization + WebSite. Secures the brand in the knowledge graph and lists official channels via sameAs.
• Service pages: Service + BreadcrumbList. Do not output FAQPage or artificial AggregateRating.
• Case studies or portfolio: Article (or TechArticle) + VideoObject (if there is a video review) + BreadcrumbList.
• Expert blog or research: Article / TechArticle linked strictly to the Person author.

JSON-LD for a B2B organization

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Organization",
  "@id": "https://example.com/#organization",
  "name": "Example Ltd",
  "legalName": "Example Ltd",
  "url": "https://example.com",
  "logo": "https://example.com/brand/logo.png",
  "telephone": "+49-30-0000-0000",
  "email": "contact@example.com",
  "sameAs": [
    "https://t.me/example_channel",
    "https://github.com/example-org"
  ],
  "contactPoint": {
    "@type": "ContactPoint",
    "telephone": "+49-30-0000-0000",
    "contactType": "sales",
    "availableLanguage": ["English", "German"]
  }
}
</script>

JSON-LD for a B2B service landing page

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "BreadcrumbList",
      "itemListElement": [
        { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://example.com" },
        { "@type": "ListItem", "position": 2, "name": "Services", "item": "https://example.com/services" },
        { "@type": "ListItem", "position": 3, "name": "Web Development", "item": "https://example.com/services/web-development" }
      ]
    },
    {
      "@type": "Service",
      "name": "High-load Web Development",
      "description": "Engineering resilient platforms on Next.js designed for peak loads.",
      "provider": {
        "@id": "https://example.com/#organization"
      },
      "areaServed": "DE",
      "serviceType": "Web Development"
    }
  ]
}
</script>

Model 2: E-commerce

(Large catalog, product grids, transactional intent)

Structured data map

• Product page: Product + Offer + real AggregateRating + BreadcrumbList. Mandatory: shippingDetails and hasMerchantReturnPolicy.
• Category listing: CollectionPage + ItemList + BreadcrumbList.

JSON-LD for a product page

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Wireless Mechanical Keyboard Keychron K2 Pro",
  "image": [
    "https://example.com/photos/1x1/photo.jpg",
    "https://example.com/photos/4x3/photo.jpg"
  ],
  "description": "Wireless custom mechanical keyboard with QMK/VIA support and RGB backlight.",
  "sku": "K2P-Q1",
  "mpn": "920-008012",
  "brand": {
    "@type": "Brand",
    "name": "Keychron"
  },
  "aggregateRating": {
    "@type": "AggregateRating",
    "ratingValue": "4.8",
    "reviewCount": "89"
  },
  "offers": {
    "@type": "Offer",
    "url": "https://example.com/keyboard-k2-pro",
    "priceCurrency": "EUR",
    "price": "119",
    "priceValidUntil": "2027-12-31",
    "itemCondition": "https://schema.org/NewCondition",
    "availability": "https://schema.org/InStock",
    "seller": {
      "@type": "Organization",
      "name": "SuperShop"
    },
    "shippingDetails": {
      "@type": "OfferShippingDetails",
      "shippingRate": {
        "@type": "MonetaryAmount",
        "value": "0",
        "currency": "EUR"
      },
      "shippingDestination": {
        "@type": "DefinedRegion",
        "addressCountry": "DE"
      },
      "deliveryTime": {
        "@type": "ShippingDeliveryTime",
        "transitTime": {
          "@type": "QuantitativeValue",
          "minValue": 1,
          "maxValue": 3,
          "unitCode": "DAY"
        }
      }
    },
    "hasMerchantReturnPolicy": {
      "@type": "MerchantReturnPolicy",
      "applicableCountry": "DE",
      "returnPolicyCategory": "https://schema.org/MerchantReturnFiniteReturnWindow",
      "merchantReturnDays": 14,
      "returnMethod": "https://schema.org/ReturnByMail",
      "returnFees": "https://schema.org/FreeReturn"
    }
  }
}
</script>

Model 3: Local business or offline services

(Geo-binding, physical address, appointments)

Structured data map

• Site branch: Specialized subtype: MedicalClinic, AutoRepair, or Restaurant. Defined with address, geo, telephone, and openingHoursSpecification.
• Specialists: Physician or Person with a worksFor connection.
• Verification: Verification in Google Business Profile is the primary factor. On-site markup validates NAP (Name, Address, Phone) consistency.

JSON-LD for a medical clinic

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "MedicalClinic",
  "@id": "https://clinic-example.com/#clinic",
  "name": "Health Plus Medical Center",
  "image": "https://clinic-example.com/images/facade.jpg",
  "telephone": "+49-30-123-45-67",
  "priceRange": "$$",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "Friedrichstraße 42",
    "addressLocality": "Berlin",
    "postalCode": "10117",
    "addressCountry": "DE"
  },
  "geo": {
    "@type": "GeoCoordinates",
    "latitude": "52.5186",
    "longitude": "13.3888"
  },
  "openingHoursSpecification": [
    {
      "@type": "OpeningHoursSpecification",
      "dayOfWeek": ["Monday", "Tuesday", "Wednesday", "Thursday", "Friday"],
      "opens": "08:00",
      "closes": "21:00"
    },
    {
      "@type": "OpeningHoursSpecification",
      "dayOfWeek": ["Saturday", "Sunday"],
      "opens": "09:00",
      "closes": "18:00"
    }
  ],
  "medicalSpecialty": [
    "Cardiovascular",
    "Dentistry",
    "Pediatric"
  ]
}
</script>

Model 4: Expert media or content portal

(Articles, analytics, news)

Structured data map

• Publications: Article, NewsArticle, or TechArticle. Must specify datePublished, dateModified, an image (≥1200px), and a full author object.
• Author pages: ProfilePage + Person. Links to social networks (sameAs), employer (worksFor), and job title (jobTitle).
• Video materials: VideoObject (with key episodes via hasPart / Clip).
• Discussions: DiscussionForumPosting (supported by Google for forums).
• Open Graph: Essential for social networks (og:title, og:image, og:description).

JSON-LD for an expert article with E-E-A-T author markup

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "TechArticle",
  "headline": "LLM Integration Architecture for Corporate Systems",
  "description": "A practical guide to deploying local language models in closed enterprise systems with data protection.",
  "image": "https://media.example.com/covers/llm-erp-1200x720.jpg",
  "datePublished": "2026-03-15T09:00:00+01:00",
  "dateModified": "2026-08-20T14:30:00+01:00",
  "mainEntityOfPage": {
    "@type": "WebPage",
    "@id": "https://media.example.com/articles/llm-erp-integration"
  },
  "author": {
    "@type": "Person",
    "name": "Alex Smith",
    "jobTitle": "Lead AI Architect",
    "url": "https://media.example.com/authors/alex-smith",
    "sameAs": [
      "https://github.com/asmith-ai"
    ]
  },
  "publisher": {
    "@type": "Organization",
    "name": "TechPulse Media",
    "logo": {
      "@type": "ImageObject",
      "url": "https://media.example.com/logo.png"
    }
  }
}
</script>

Antipatterns

1. Fake AggregateRating on service pages: Marking up a 4.9–5.0 rating with 150 reviews on a service page when no client reviews exist. Google applies manual actions for spammy structured markup. The page loses all snippets, including breadcrumbs, and falls under algorithmic spam filtering suspicion.
2. Hidden data markup: Passing prices, questions, lists, or text in JSON-LD that are not in the visual DOM tree (or hidden via display: none for keyword stuffing). Leads to penalization for attempting to manipulate search bots.
3. Dead FAQPage on commercial landing pages: Wasting resources generating complex FAQ schemas on landing pages expecting interactive snippets. The snippet will not appear due to deprecations.
4. Mixing Microdata and JSON-LD in one template: Marking up part of the fields with HTML tags (itemscope, itemprop) and part with JSON-LD script. Causes entity duplication, parser errors, and validation breakage during layout refactoring. A unified JSON-LD block is the de facto standard.

Audit checklist

• 1. Implementation format: All structured data is output in JSON-LD inside a <script type="application/ld+json"> tag.
• 2. SSR validity: The markup is server-side rendered and present in the raw HTML. The bot does not need to execute heavy JavaScript to read the schema.
• 3. Breadcrumbs (BreadcrumbList): Implemented on all pages deeper than the homepage. All intermediate links contain absolute URLs.
• 4. Brand and organization (Organization): The homepage contains full contact details and a sameAs array with links to authoritative external company profiles.
• 5. E-commerce: Cards are marked up with the Product schema, including shippingDetails and hasMerchantReturnPolicy blocks.
• 6. Local business: Profiles in Google Business Profile are synchronized with the site for NAP consistency.
• 7. Verification before deployment: Validation via the Google Rich Results Test yields zero errors and valid rich results.