Google has been quietly trimming its rich results programme for a few years now, and the cuts accelerated through 2025 and into 2026. FAQ rich results went first, then How-To snippets in desktop search, then a handful of others that many sites had been relying on for click-through boosts. If you've noticed your star ratings or expanded snippets disappearing from the search results page, you are not imagining it.
The question worth asking now is not "what did Google remove?" but "what still works, and is it worth the investment?" The answer is more nuanced than most blog posts let on. Some rich result types remain genuinely powerful. Others were always marginal and their removal barely matters. A few schema types that Google no longer surfaces as visual rich results have found a second life feeding AI search engines like ChatGPT, Perplexity, and Gemini. That last point is where the real opportunity sits in 2026.
What Google Actually Cut (and Why It Matters)
Google's stated reason for removing FAQ and How-To rich results from most surfaces was that they were being abused. Pages were stuffing FAQs with keyword-rich questions that served no real user need, and How-To markup was appearing on pages that contained nothing resembling step-by-step instructions. The results page became visually noisy without adding value, so Google pulled back.
The broader 2025 - 2026 pattern has been a move away from rich results that expand the visual footprint of any single result. Google is protecting real estate for ads, AI Overviews, and its own knowledge panels. That context matters because it tells you something about which rich results have survived: the ones that serve the user's intent directly at the moment of the search, not the ones that just made your blue link look bigger.
Here is what was removed or significantly reduced:
- FAQ rich results - removed for most sites in 2023, fully deprecated for standard results by 2025. Government and health sites retained limited support for a time.
- How-To rich results - removed from desktop, then pulled from mobile as well.
- Product snippets with seller ratings - display rules tightened significantly; many e-commerce sites lost aggregate star ratings on non-Shopping tab results.
- Video carousel prominence - reduced on informational queries where AI Overviews now dominate.
Worth noting: FAQ schema is not dead even though the visual rich result is gone. That distinction is important and we will come back to it.
The Rich Result Types That Still Perform
Product Schema and Merchant Centre Integration
If you run an e-commerce store, Product schema remains one of the highest-return structured data investments you can make. The visual treatment in Google Shopping results, including price, availability, and review stars, is still very much alive. The key change is that Google now places significantly more weight on your Merchant Centre feed being consistent with your on-page Product markup. Discrepancies get your rich results suppressed quickly.
For Shopify merchants in particular, the default Product schema generated by most themes is incomplete. It typically omits offers with accurate priceCurrency, availability, and condition values. Fixing that alone can restore rich results that have been silently suppressed.
Offer schema deserves its own focus. The Offer schema type now feeds directly into AI shopping recommendations in Perplexity and Google's AI Mode, not just traditional rich results. That dual value makes it one of the most important schema investments you can make right now.
Review and Aggregate Rating Schema
Star ratings in search results are still live for several content types: recipes, products, software applications, books, courses, and local businesses. The rules have tightened. Google requires that reviews be first-party (collected and displayed by you) or from an approved third-party aggregator, and the schema must accurately reflect what is on the page.
The click-through impact of star ratings is real and well-documented. A result with 4.7 stars and 230 reviews displayed in the snippet will consistently outperform the same result without them, often by 15 - 30% on commercial queries. That advantage has not gone away.
Recipe Schema
Recipe rich results are one of the most stable categories. Google has not pulled back on these, and the visual treatment is generous: image, cook time, rating, calorie count, and sometimes video. If your site publishes recipe content, this is low-hanging fruit and should be fully implemented.
Video Schema
Video rich results, particularly on YouTube-hosted content embedded on your own pages, still appear with meaningful frequency. The VideoObject schema type helps Google understand duration, upload date, description, and thumbnail. The key metric Google uses to decide whether to show the rich result is whether the video is genuinely the primary content of the page, not just a decorative embed.
Event Schema
Event markup continues to generate strong rich results in search, including date, location, ticket price, and availability. For businesses running live or online events, this is worth implementing properly. Google's event rich results appear prominently on mobile in particular, and the structured data also feeds into event discovery in AI answers.
Local Business and Organisation Schema
This category never generated a traditional "rich result" in the snippet sense, but it feeds the Knowledge Panel that appears for branded searches and increasingly informs what AI models say about your business. Getting your LocalBusiness, Organization, address, telephone, and openingHours properties accurate and complete is foundational. It is not glamorous but the downstream effects on AI citation and knowledge panel accuracy are significant.
The Schema That No Longer Gets Rich Results But Still Matters
This is the part most articles miss entirely. When Google removes a rich result type, it does not mean the schema becomes useless. It means the schema stops generating a visual treatment in traditional search. The structured data still exists in your page, and AI crawlers, Perplexity's indexer, Bing's AI pipeline, and others read it directly.
FAQ schema is the clearest example. Google dropped the visual accordion in search results. But when you ask ChatGPT or Perplexity a question, and a page with well-structured FAQ schema answers that question cleanly, the AI model is far more likely to cite that page and quote from it accurately. The machine-readable format makes it easier for the model to extract the right answer. So the schema that "failed" in Google still wins in AI search.
The same logic applies to HowTo, WebPage, and AboutPage schema. None of these generate a flashy visual rich result in Google. All of them improve the clarity and structure of your content for AI systems that are increasingly the first point of contact between a potential customer and information about your business.
This is the core of what FlinnSchema focuses on: the overlap between traditional schema best practice and what AI search engines actually need to cite, recommend, and trust your content. The two goals are more aligned than most people realise, but the emphasis has shifted in ways that most SEO tooling has not yet caught up with.
How to Decide What to Implement First
Prioritisation depends on your business type, but here is a practical framework:
For E-commerce Sites
Start with Product and Offer schema. Make sure every product page has accurate price, availability, and condition data. Add AggregateRating where you have genuine reviews. Then layer in BreadcrumbList for navigation and Organisation at site level. This combination covers the highest-value Google rich results and simultaneously feeds AI shopping recommendations.
For Content and Media Sites
Article, VideoObject, and Recipe schema (where applicable) should be your foundation. Add BreadcrumbList and WebPage schema to every page. If you publish event coverage or interviews, EventSchema and Person schema add value. For AI visibility specifically, well-structured FAQ content on every major topic page is worth adding even without the rich result trigger.
For Service Businesses and SaaS
Organisation and LocalBusiness schema are non-negotiable. Add Service schema for each service offering, SoftwareApplication schema if you have a product, and FAQ schema on your key landing pages. The SoftwareApplication schema type in particular is underused and drives meaningful AI citations for SaaS products. Review your AboutPage schema as well, since AI models use it to form initial impressions of your brand's credibility and scope.
Testing and Monitoring Your Rich Results
Google's Rich Results Test tool remains the fastest way to validate your markup. Run it on your key page templates, not just your homepage. The most common issues in 2026 are:
- Missing required properties (particularly
offerson Product pages) - Inconsistency between schema values and visible page content (Google will suppress the rich result)
- Using deprecated schema types that Google no longer supports
- Implementing schema via third-party plugins that generate outdated or incorrect property names
Beyond Google's own tool, monitoring your actual appearance in search requires checking Google Search Console under the "Enhancements" section. This shows you which rich result types are being detected across your site and which have errors or warnings. If you are serious about tracking AI visibility alongside traditional rich results, tools like Profound, Otterly, and Peec now offer share-of-voice tracking across AI answer engines. There is a detailed comparison of those tools if you want to dig into which one fits your needs.
The overall picture in 2026 is this: Google's rich results programme is smaller and more selective than it was three years ago. The sites that are winning are the ones that implement schema correctly for the types that still work, and that understand structured data as a signal for AI systems, not just a trick for getting bigger snippets. Those are different goals that happen to require much of the same work.
Frequently Asked Questions
Is it still worth adding FAQ schema if Google doesn't show FAQ rich results?
Yes. FAQ schema structures your content in a way that AI search engines like ChatGPT, Perplexity, and Gemini can parse directly. Even without a visual rich result in Google, the markup helps AI models extract accurate answers and cite your page. For informational content, that AI citation is increasingly more valuable than the Google rich result ever was.
Which schema types are most likely to generate rich results in Google right now?
The most reliably active rich result types in 2026 are Product (with Offer and AggregateRating), Recipe, Event, Video, and Review. Local Business schema feeds Knowledge Panels rather than traditional snippets but is equally important for branded searches. BreadcrumbList schema is also widely supported and improves your URL display in results.
Does incorrect schema hurt your rankings?
Incorrect schema generally causes Google to ignore or suppress the rich result rather than penalise the page. However, if your schema is deliberately misleading, for example showing a five-star rating that is not reflected on the page, Google can issue a manual action. The practical risk for most sites is not a penalty but simply losing the rich result treatment you were hoping for.
How do I know if my schema is actually driving clicks?
Google Search Console is your starting point. The Enhancements reports show detected rich results, and the Performance report lets you filter by page and query to see click-through rates. If you implement schema on a set of pages and compare click-through before and after, over a period of at least four to six weeks, you will start to see whether the rich result is making a measurable difference. For AI visibility, you need a dedicated monitoring tool since Search Console does not track citations in AI answers.