Schema markup in 2026 is the single most underused free SEO lever most sites have. The right schema gets you rich results in Google, reliable retrieval by AI search engines, and clean entity associations that compound over time. The wrong schema gets ignored or, worse, flagged as spam. This guide is the practical walkthrough — every type that matters for marketing sites and SaaS, what each one does, when to use it, and the validation steps that catch the bugs that cost rich-result eligibility.

Why schema matters more in 2026

  • Rich results still drive 18–35% higher click-through rates than plain blue links for queries where they appear (FAQ, How-to, Product, Article-with-byline).
  • AI search retrieval uses structured data heavily. ChatGPT, Perplexity, Claude with browsing, and Google AI Overviews all weight schema-marked content as more reliable for citation. See our AEO & GEO playbook for why this matters.
  • Entity association. Schema explicitly tells Google what the page is about, who's behind it, and how it relates to other entities. The signal compounds across crawls.
  • Free, durable, hard to copy. Schema discipline is a moat against competitors who don't implement it.

JSON-LD is the format to use

Three schema formats exist (Microdata, RDFa, JSON-LD). In 2026 use JSON-LD exclusively. It lives in <script type="application/ld+json">blocks separate from your HTML markup, doesn't couple schema to visible elements, and is recommended by Google. The other two are legacy.

Schema types every marketing site should ship

Organization (sitewide)

Tells Google who runs the site. Goes on the homepage at minimum; we ship it on every page via the layout. Include legal name, logo URL, address, contact, social profiles, and founding date.

WebSite + SearchAction (sitewide)

Lets Google show a sitelink search box in SERPs for your brand. Pair with the Organization schema. Include the search URL template even if your site search is basic.

Article (every blog post)

Article schema enables rich results and AI search citations. Required: headline, description, datePublished, author, publisher, mainEntityOfPage. Recommended: dateModified, image, keywords, articleSection.

Tells Google how the page sits in the site hierarchy. Surfaces as breadcrumbs in SERPs. Implementation is mechanical: list the path from home to the current page, with name + item URL for each.

FAQPage (where it makes sense)

Surfaces questions and answers directly in Google SERPs (as collapsible Q&A or, increasingly, as expanded passages). Use it when the page genuinely has FAQ content. Don't add FAQPage schema to invented FAQs. Google has gotten stricter; misuse triggers manual penalties.

HowTo (genuine instructional content only)

Steps + tools + time required. Useful for genuine how-to content, less rewarded by Google than it was in 2022 (the rich result is now suppressed in many categories). Implement where appropriate but don't expect the rich-result placement.

E-commerce schema

Product

Required for product rich results. Include name, image, description, sku, brand, offers (with price, priceCurrency, availability, url), aggregateRating, review.

Pricing accuracy matters. If your structured price doesn't match your visible price, Google flags it. Common cause: caching issues where the schema is stale vs the page.

Offer

Nested inside Product. Include price, priceCurrency, priceValidUntil, availability (InStock, OutOfStock, PreOrder, etc.), itemCondition, shippingDetails, hasMerchantReturnPolicy. Modern Google requires shipping details and return policy for full eligibility.

Review and AggregateRating

Star ratings in SERP. Include reviewCount, ratingValue, bestRating, and at least one Review item nested. Reviews must be genuine. Google checks; fake reviews trigger penalties.

Local business schema

LocalBusiness (and subtypes)

Required if you serve a physical location or service area. Include name, address, telephone, openingHours, priceRange, geo coordinates, sameAs links to social, image. Use specific subtypes when applicable: Dentist, LegalService, Restaurant, AutoRepair, GeneralContractor, etc.

Service (with serviceArea)

For service businesses, complement LocalBusiness with one Service entity per major offering. Include name, description, provider, serviceArea, offers, areaServed.

People and authorship

Person

For author bios, leadership pages, coach/consultant sites. Include name, jobTitle, image, sameAs (LinkedIn, X, etc.), knowsAbout, worksFor.

Person schema feeds entity recognition in AI search. The clearer your Person markup, the more likely ChatGPT or Perplexity surfaces you for queries about your domain expertise.

Author byline on Articles

Article schema supports author as a Person object. Use full Person nesting (not just a string name) so the author entity is properly recognized. Connects authorship to expertise across the site.

Case studies and project work

CreativeWork

For agency case studies and portfolio entries. Include name, headline, description, image, creator (your org), datePublished, keywords. We added this to every case study page on this site after the SEO audit identified it as missing.

SaaS-specific schema

SoftwareApplication for product pages. Include name, operatingSystem, applicationCategory, offers (with price), aggregateRating. Useful for SaaS pricing pages and product review submissions.

Validation and testing

Three tools matter:

  • Google Rich Results Test (https://search.google.com/test/rich-results) — checks eligibility for rich result types per page.
  • Schema Markup Validator (https://validator.schema.org/) — generic validator covering all schema.org types, not just Google's subset.
  • Search Console > Enhancements — shows rich results detected on your live site, with errors and warnings.

Run all three. Each catches different bugs. The Rich Results Test in particular shows you exactly what Google sees, which isn't always what you intended.

A practical implementation pattern

How we ship schema across the marketing sites we build:

  1. Sitewide layer: Organization + WebSite in the root layout. Renders on every page.
  2. Page-type layer: Article, Product, LocalBusiness, etc. attached to specific page templates.
  3. Per-instance fields: Pulled from the content source (CMS, database, code-as-content).
  4. Validation in CI: A test suite that runs the Rich Results Test API against every page before deploy. Catches regressions.

Schema mistakes that cost rich results

  • Marking up content that isn't visible. Google requires schema content to match visible page content. Hidden FAQ schema, fake testimonials, prices that don't appear on the page — all triggers.
  • Inflated review counts. 5,000 5-star reviews on a site with 30 indexed pages is suspicious. Google checks plausibility.
  • FAQPage schema on every page. Adding 6 invented FAQs to every product page. Manual penalty territory.
  • Stale prices in Offer schema. When schema updates lag behind actual price changes, listings disappear from Shopping until reconciled.
  • Missing required fields. Article without headline, Product without offers, LocalBusiness without address. Schema validates but doesn't qualify for rich results.

Want a schema audit?

Send us your URL. We'll send a one-page report on your current schema coverage, the rich-result types you're eligible (or not eligible) for, and the highest-leverage additions — within three working days.

Book a consultation →

Frequently asked questions

Which schema format should I use in 2026 — Microdata, RDFa, or JSON-LD?

JSON-LD exclusively. It&apos;s recommended by Google, lives in script tags separate from your HTML, and doesn&apos;t couple schema to visible elements. Microdata and RDFa are legacy. Most modern frameworks (Next.js, Astro, Nuxt) and CMSes support JSON-LD natively.

Does FAQPage schema still produce rich results in 2026?

Yes, but Google has tightened eligibility. FAQ rich results show only when the page genuinely has FAQ content visible, the questions are non-spammy, and the site has authority for the topic. Adding FAQPage schema to every product page or inventing 6 FAQs to qualify is now a manual-penalty trigger. Use FAQ schema when there&apos;s real FAQ content on the page; skip it otherwise.

Do I need Product schema if I&apos;m not on Google Shopping?

Yes, for several reasons. Product schema enables organic SERP rich results (price, availability, ratings) even outside Shopping. AI search engines (ChatGPT, Perplexity) heavily use Product schema for retrieval. Shopify, WooCommerce, and most modern e-commerce platforms ship Product schema by default — verify yours does and that the data matches the visible page.

Should I add LocalBusiness schema if I&apos;m not local?

No. LocalBusiness is for businesses with a physical location or specific service area. Misusing it to chase local-pack visibility on a non-local business is detected and ignored. Use Organization schema instead for a national or global brand.

How do I validate my schema in 2026?

Three tools, run all three: Google&apos;s Rich Results Test (search.google.com/test/rich-results) for rich-result eligibility per page; Schema Markup Validator (validator.schema.org) for generic validation across all schema.org types; and Search Console&apos;s Enhancements report for what Google detects on your live site. Each catches different bugs. Add automated schema validation to your CI pipeline if you ship frequently.

Which schema types matter most for AI search retrieval?

Article and FAQPage are the most consistently retrieved by ChatGPT, Perplexity, Claude with browsing, and Google AI Overviews in 2026. Person schema (with knowsAbout) helps for individual coaches and consultants who want to be retrievable for their domain. Product schema is essential for any e-commerce or SaaS pricing page. Organization helps with entity recognition across the site.