When Google dropped FAQPage and HowTo from its rich results in 2023, a lot of sites ripped the schema out entirely. That was a mistake. The rich result went away. The schema is still being parsed, still being used for entity grounding, and still being lifted by AI engines into citations.

The right read on the FAQPage deprecation: Google stopped showing the rich snippet in search results. Google did not stop parsing the markup, did not stop indexing the structured Q&A, and did not stop using it as a hint about page topic. Other consumers of schema (Bing, AI engines, voice assistants, internal Google systems) never stopped using it at all.

The practical question is: which schema types currently earn placements that drive real attention, and which are just markup tax that nobody is consuming? Here is the honest breakdown for 2026.

Schema types that still earn placements

These are the types where the markup produces a visible, traffic-driving placement (rich result, knowledge panel, AI citation hint, or platform feature) in 2026.

Article (and NewsArticle, BlogPosting)

Article schema is the most underrated workhorse in 2026. It powers the Top Stories carousel, the date-confirmed result format, the AI Overview citation grounding, and the Discover surface on mobile. Every editorial page should have it. The fields that matter most are headline, datePublished, dateModified, author (with proper Person or Organization object), and image.

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Schema markup that still matters after the FAQPage deprecation",
  "datePublished": "2026-04-21",
  "dateModified": "2026-04-21",
  "author": {
    "@type": "Organization",
    "name": "ConceptSEO",
    "url": "https://seo.concept211.com"
  },
  "publisher": {
    "@type": "Organization",
    "name": "ConceptSEO",
    "logo": {
      "@type": "ImageObject",
      "url": "https://seo.concept211.com/icons/ConceptSEO.png"
    }
  },
  "image": "https://seo.concept211.com/og/article.png"
}
</script>

Product

Product schema with price, availability, and review aggregate is mandatory for any commerce page that wants to appear in product carousels, shopping graph results, and AI shopping comparisons. The merchant listing experiences (in Google Search and Google Shopping) lean entirely on this markup. Skip it on a product page and that page is invisible to half the shopping query surface.

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Example Product",
  "image": "https://example.com/img.jpg",
  "description": "...",
  "sku": "SKU-001",
  "brand": { "@type": "Brand", "name": "Example" },
  "offers": {
    "@type": "Offer",
    "url": "https://example.com/product",
    "priceCurrency": "USD",
    "price": "29.99",
    "availability": "https://schema.org/InStock"
  },
  "aggregateRating": {
    "@type": "AggregateRating",
    "ratingValue": "4.6",
    "reviewCount": "127"
  }
}
</script>
Hands typing on a laptop with code displayed on screen

Photo by TREEDEO.ST on Pexels

LocalBusiness

LocalBusiness schema is the on-site complement to Google Business Profile. It does not replace GBP, but it provides the entity grounding that links your website to your physical location for AI engines that cannot reliably parse GBP directly. Use the most specific subtype available (Restaurant, Dentist, Plumber, LawFirm) rather than the generic LocalBusiness type.

The fields that matter: name, address (full PostalAddress object with all fields), telephone, openingHoursSpecification, geo (lat/lng), priceRange, and sameAs pointing to your GBP listing URL.

BreadcrumbList

Still rendered in SERP as the breadcrumb display path under the title. Trivially cheap to add, modest but real CTR lift, near-universal support. There is no reason a content site or e-commerce site in 2026 should not have BreadcrumbList on every non-homepage URL.

Organization (with sameAs)

The single most leveraged schema type for GEO and brand-mention SEO. Put it once in the site-wide layout. Include logo, url, contactPoint, and the full sameAs array pointing to every authoritative external profile: Wikidata, Wikipedia, Crunchbase, LinkedIn, Twitter/X, Facebook, GitHub, YouTube, GBP. This is how AI engines disambiguate your brand from any other entity with a similar name.

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Example Corp",
  "url": "https://example.com",
  "logo": "https://example.com/logo.png",
  "sameAs": [
    "https://www.wikidata.org/wiki/Q12345678",
    "https://en.wikipedia.org/wiki/Example_Corp",
    "https://www.linkedin.com/company/example-corp",
    "https://www.crunchbase.com/organization/example-corp",
    "https://twitter.com/example"
  ]
}
</script>

Person

For authored content, individual founder pages, expert profiles, and any page where personal authority is the credibility signal. Person schema lets AI engines build an author entity that connects across all your published work. With sameAs to a Wikidata QID, this is the cleanest way to establish authorial expertise that travels with the person across sites.

Recipe

Still produces the dedicated Recipe rich result with ratings, time, and ingredients. Voice assistants lean heavily on Recipe markup for cooking-related queries. If you publish recipes and do not have this markup, you are invisible to roughly 40% of recipe-driven traffic surfaces.

VideoObject

Powers the video carousel placements in SERP, the key moments feature, and the video preview on hover. Required for YouTube embed pages, original hosted video, and video-heavy product pages. The fields that matter: name, description, thumbnailUrl, uploadDate, duration, and contentUrl (or embedUrl).

Event

For anything date and location bound: concerts, conferences, classes, webinars. Powers the dedicated event card placement in SERP and Google's events vertical. Include startDate, endDate, location (Place object with address), eventStatus, and eventAttendanceMode (online, offline, or mixed). Without eventAttendanceMode the listing often gets discounted.

Close-up of HTML and JavaScript code on a developer's screen

Photo by Antonio Batinić on Pexels

Schema types where the rich result went away but the markup still matters

This is where most of the bad advice has been concentrated. The rich result deprecation is real but narrow. The markup continues to do work.

FAQPage

Lost the rich snippet in SERP. Still parsed by Google. Still consumed by AI Overviews and Perplexity for question-answer grounding. Still used by voice assistants. Still useful for marking up genuine Q&A on product pages, documentation, and support content. The cost to add is near zero. Remove it only if it was a vehicle for spam Q&A. Keep it for legitimate question-driven content.

HowTo

Lost the rich snippet. Still parsed. Still used by AI engines for procedural answer grounding. If you have step-by-step instructions, mark them up. The user-facing impact is gone but the AI-facing impact is intact.

Schema types that genuinely do not move the needle anymore

A few types either never produced placements or have lost meaningful support. Worth knowing so you do not spend time on them.

The rule for deciding what to mark up

For every page, ask three questions:

  1. Does this content match a schema type that has either a visible Google rich result or known AI-engine consumption? If yes, mark it up.
  2. Is the markup load-bearing for entity disambiguation (Organization, Person, LocalBusiness, sameAs)? If yes, mark it up regardless of rich result status.
  3. Is this an editorial page that could be cited in AI Overviews? If yes, mark it up as Article with author and dates, regardless of whether it produces a rich result today.

If none of the three answers are yes, skip the schema. Bloating pages with markup that nobody consumes is a real cost (longer HTML, slower parse, more risk of validation errors) for zero benefit.

Quotable
The FAQPage and HowTo deprecation only removed the visible rich result in Google Search. Google did not stop parsing the markup, AI engines did not stop using it for question-answer grounding, and voice assistants did not stop relying on it for procedural answers. The schema types that still earn visible placements in 2026 are Article, Product, LocalBusiness, BreadcrumbList, Organization with a populated sameAs array, Person, Recipe, VideoObject, and Event. Organization schema with sameAs pointing to Wikidata, Wikipedia, LinkedIn, and Crunchbase is the highest-leverage single piece of markup any brand can add, because it gives every AI engine the entity edges they need to know who you are. Sites that ripped out FAQPage after the rich result was removed silently degraded their AI citation rate without realizing it. Sites that kept it and added Organization sameAs picked up citation surface they did not have before. The schema layer is not just about SERP features anymore. It is the connective tissue between your site and every machine that reads the web.

What to ship this quarter

If your schema layer has been neglected, here is the minimum viable upgrade, in priority order:

  1. Site-wide Organization schema with logo, url, and the full sameAs array. Put it in your layout, ship once.
  2. Article schema on every editorial page with headline, datePublished, dateModified, author, publisher, image.
  3. BreadcrumbList on every non-homepage URL.
  4. Product schema on every commerce page with price, availability, aggregateRating.
  5. LocalBusiness on the contact page (and on location-specific landing pages if multi-location).
  6. Person schema on author bio pages with sameAs to professional profiles.

Validate everything with the Schema Markup Validator (schema.org/validator) and Google's Rich Results Test. Fix the errors, ignore the warnings about optional fields you do not have, and ship. Then keep going.