Schema.org defines more than 800 types and thousands of properties. Google supports only a small subset for rich results, and the support is uneven: some types produce visible features in the search results, some give Google context it may use internally, and most are read and then ignored. If you have ever marked up an entire site and watched Search Console report zero rich results, the cause is usually the same. Effort went into types Google does not surface. Knowing which types actually matter is the difference between markup that changes how your listing looks and markup that changes nothing.
The mismatch exists because schema.org is a general vocabulary for describing anything on the web, while Google’s rich results system is a narrow feature set built around specific SERP treatments: rating stars, recipe cards, product prices, event listings, job postings, and a handful of others. Types outside that set can be valid markup and still produce no visible result. The list is not fixed either. Google has been trimming it aggressively, so a priority plan written a few years ago is now partly wrong.
Which schema types still produce visible rich results #
Google’s structured data documentation lists the types eligible for rich results. As of 2026, after two rounds of pruning, the durable, broadly useful ones are:
| Schema type | Rich result | Where it matters |
|---|---|---|
| Product (with Offer, AggregateRating) | Price, availability, rating stars | Ecommerce; appears in standard and shopping surfaces |
| Review / AggregateRating | Inline star ratings | Visible across many result types |
| Recipe | Recipe cards with image, time, rating | Cooking and food sites |
| Article / NewsArticle | Top stories, news carousels | News and publishing |
| VideoObject | Video thumbnails, key moments | Video content |
| Event | Listings with date and location | Ticketing, venues, event sites |
| JobPosting | Google for Jobs integration | Job boards and employers |
| LocalBusiness | Local panel and map details | Local SEO |
| Organization | Knowledge panel for brand queries | Brand recognition |
| BreadcrumbList | Breadcrumb trail in the listing | Universal small lift |
| ProfilePage | Author and creator attribution in results | First-hand-perspective content |
| DiscussionForumPosting | Forum threads and Q&A surfaced in results | Community and support forums |
Two of those, ProfilePage and DiscussionForumPosting, are relatively recent. Google introduced both in late 2023 for content built around named authors and community discussion, and they remain supported where they genuinely describe the page.
The types Google has retired, and why they matter to your plan #
This is where older schema advice, including advice you may have implemented, goes wrong. Google has removed several rich result types outright.
HowTo rich results were the first major casualty. Google stopped showing them on desktop in September 2023 and finished the removal on mobile in 2024, per its Search Central announcement. Any guide that still describes HowTo markup as a live way to win step-by-step SERP displays is describing a feature that is gone.
FAQ rich results followed a longer path to the same place. In August 2023 Google restricted them to well-known, authoritative government and health sites, cutting off almost everyone else. Then on May 7, 2026 it removed FAQ rich results entirely, including for the government and health sites that had kept them. FAQ is now fully deprecated. If a resource tells you FAQPage markup still earns expandable Q&A snippets for health or government pages, that is out of date as of mid-2026.
A larger cut landed in between. In its June 12, 2025 announcement about simplifying the search results page, Google dropped seven more types at once: Book Actions, Course Info, Claim Review, Estimated Salary, Learning Video, Special Announcement, and Vehicle Listing. Course Info and Vehicle Listing are worth flagging because they were promoted heavily at launch, so plenty of 2023 and 2024 guidance still treats them as growth opportunities. They no longer produce rich results.
Retired types do not hurt you if the markup stays on the page, since Google ignores unused structured data rather than penalizing it. But they do nothing, so continuing to invest engineering time in them is wasted. The retirements are the strongest argument for treating schema as something you review on a schedule rather than ship once.
The types Google reads for context, and the ones it ignores #
A second group produces no visible feature yet still helps Google understand the page: Person properties on author bylines feed its assessment of expertise and authorship, Organization properties on publisher information support attribution and authority signals, WebPage with author and date properties provides freshness context, and ImageObject properties assist image search and licensing display. None of these transform your listing, and their effect is indirect and hard to isolate. Mark them up accurately if they fit, but do not expect a visible payoff.
Most of schema.org falls into a third group Google reads and ignores: niche CreativeWork subtypes, Action types for user interactions, and generic Service descriptions produce no features and show no detectable ranking influence. Marking them up is harmless, because Google ignores rather than penalizes, but the time is better spent on supported types. The urge to mark up everything is exactly what produces a Search Console dashboard full of valid schema and no rich results.
How to set priority for a real site #
For a site starting from scratch, order the work by what Google will actually show.
First, the types that match your primary content, because this is where visible results come from. Ecommerce does Product with Offer, AggregateRating, and Review. News does NewsArticle with Person author and Organization publisher. A recipe site does Recipe with rating, time, and nutrition. A service business does LocalBusiness with hours, address, and services. A job board does JobPosting with every required property.
Second, the types that apply to nearly any site: Organization on the home page or footer, BreadcrumbList wherever breadcrumb navigation exists, and WebPage with proper author and date properties.
Third, the types that add features beyond your core content, such as VideoObject where video appears on important pages, or Event for time-bounded content. Everything else is a fourth tier where the cost-benefit is unclear and, for most sites, optional. Notice what is no longer on the list: HowTo, FAQPage, Course Info, and Vehicle Listing all lost their rich results, so none belongs in a priority plan aimed at visible features.
The implementation mistakes that recur in audits #
- Marking up content users cannot see. Schema must describe what is on the page; reviews in markup with none visible is a manual action trigger, not a shortcut.
- AggregateRating without real reviews. The aggregate must reflect actual review data shown on the page. Invented ratings are one of Google’s more aggressive manual action categories.
- The wrong type, such as Article where Product is correct, or LocalBusiness where Organization is more accurate.
- Missing required properties. Product needs name and image; JobPosting needs title, hiringOrganization, and jobLocation. Validators flag these.
- JSON-LD that does not match the rendered HTML. When JavaScript changes the page after the markup renders, Google may read something different from what users see, so the schema should reflect the final rendered state.
- Recipe schema on non-recipe pages, or any speculative markup. Google’s spam systems detect mismatched schema and can reduce rich result eligibility site-wide.
- Schema on pages blocked by robots.txt. The quiet failure: perfect markup on a page Google cannot crawl produces zero rich results. Cross-check pages carrying schema against your robots.txt rules, and confirm in Search Console URL Inspection that the page is not reported as blocked.
Choosing the specific type over the general one #
Schema.org is hierarchical. Article has subtypes such as NewsArticle and BlogPosting; Organization has LocalBusiness, Corporation, and EducationalOrganization. Use the most specific type that accurately fits the content, because NewsArticle carries more signal than the parent Article, and LocalBusiness carries more than Organization. The trade-off is that specific types often have stricter required properties, so fall back to the parent when you cannot meet them. Reaching for Thing or CreativeWork when a precise type exists is the wrong instinct; Google’s processing favors specificity.
How schema connects to ranking #
Schema markup is not a direct ranking factor, and Google representatives including John Mueller have said so repeatedly. What it does is enable rich results that can raise click-through rate, provide context algorithms use to understand the page, connect entities through sameAs links, and help Google pick the right SERP feature. A product page showing a star rating tends to earn more clicks than the same page without it, and those user signals can influence ranking over time. The effect is indirect but real. Because that payoff depends on Google continuing to support a given feature, the retirements above matter directly: markup for a deprecated type cannot lift CTR because there is nothing left to display.
FAQ #
Is schema markup a ranking factor?
Not directly. Google has stated schema is not a traditional ranking signal. It works by enabling rich results and providing context, both of which can influence traffic and, indirectly, ranking over time.
Should I remove FAQPage and HowTo markup now that the rich results are gone?
There is no need to rush. Unused structured data does not cause problems for Search; it simply produces no visible result. Remove it when convenient, and do not invest new effort in it.
Which validator should I trust?
Google’s Rich Results Test shows what Google sees and which rich result types a URL is eligible for. The Schema.org Validator checks vocabulary correctness independent of Google. Search Console enhancement reports track schema health across the site over time.
How often should I review my schema?
Annually at minimum for active sites. Google removed HowTo, then seven more types in June 2025, then FAQ entirely in May 2026, so a priority plan more than a year or two old likely contains types that no longer produce anything.
The goal of schema work is not comprehensive markup of every page with every type. It is accurate, validated, prioritized markup of the types that match your primary content and that Google still supports. Start with the highest-impact type for your category, validate before you deploy, keep the markup aligned with what users see, and put the review on a calendar. The implementation details keep changing, as the last three years show. The discipline of prioritizing what Google actually uses does not.