Users click on a rich result 58% of the time, compared to just 41% for a standard result — according to a Milestone analysis of more than 4.5 million queries. Those 17 percentage points don't come from better copywriting: they come from speaking to Google in its own language, the language of structured data. But 2026 changed the rules. On May 7th, Google permanently retired the once-popular FAQ rich results. This guide explains which structured data types are still worth your time, what the key types are, and how to implement them with JSON-LD without making mistakes.
What structured data is and why it matters for SEO
Structured data is a standardized format for describing a page's content in a way that search engines understand unambiguously. Instead of leaving Google to guess whether a number is a price, a publication date, or a product rating, you tell it explicitly through a shared vocabulary.
That distinction — "Google interprets" versus "Google knows" — is the key. The visible text on a page is written for people; structured data is written for machines. When both align, the search engine can display your page with enriched elements: star ratings, breadcrumbs, expandable questions, company information, or product cards showing price and availability.
The SEO impact is not direct on rankings — structured data is not, on its own, a ranking factor — but it does have a very real indirect influence:
- Improves CTR. A result with stars, a price, or an image stands out in a crowded SERP and attracts more clicks, as confirmed by the Milestone data cited at the top of this guide.
- Clarifies context. It helps Google correctly classify your content (an article, a product, a recipe, an event), reducing misinterpretation.
- Feeds generative search. AI engines and assistants consume structured data to cite sources and construct answers, linking it directly to Generative Engine Optimization (GEO).
In short: structured data won't magically boost your rankings, but it makes the position you already hold work much harder in your favor. If you want to understand how it fits within a broader technical strategy, our technical SEO service covers semantic markup alongside performance, crawlability, and indexation.
Schema.org, JSON-LD, Microdata, and RDFa: which format to use
It's worth separating two concepts that are often confused. Schema.org is the vocabulary; JSON-LD, Microdata, and RDFa are the formats (the syntax) with which you write that vocabulary on your page.
Schema.org is a collaborative project founded by Google, Microsoft, Yahoo, and Yandex. According to the Schema.org documentation itself, its vocabulary currently includes on the order of 800 types, around 1,500 properties, plus enumerations and datatypes. It is, in practice, the shared dictionary that all major search engines agree to understand.
On top of that dictionary, you can write with three different "grammars," all accepted by Google:
| Format | Where it lives | Readability | Maintenance | Google's recommendation |
|---|---|---|---|---|
| JSON-LD | Inside a separate <script> block, not in visible HTML |
High (isolated block) | Simple and scalable | Recommended |
| Microdata | Attributes embedded in visible HTML | Low (mixed with markup) | Brittle and verbose | Accepted |
| RDFa | Attributes embedded in visible HTML | Low | Complex | Accepted |
The recommendation is clear. According to Google Search Central, JSON-LD is the recommended format for structured data because it is the easiest solution to implement and maintain at scale, and the least error-prone. The reason is practical: living inside an independent <script> block, you don't have to "pollute" every tag in your HTML or worry about breaking the markup when you redesign a template. You can generate it dynamically from the server or framework and reuse it across thousands of pages.
Microdata and RDFa are still valid, and you'll find them on older sites, but today there's no reason to start a new project with them. The rest of this guide focuses on JSON-LD.
The 5 most useful structured data types
With hundreds of types available, the temptation is to mark up everything. That's a mistake. In B2B and in most corporate sites, five types cover 90% of the real value. These are the ones we recommend prioritizing.
1. Organization
Describes your company: name, logo, URL, social profiles, and contact details. It's the foundation of brand identity for Google and feeds the knowledge panel. Usually placed on the homepage or in a global layout so it appears across the entire site.
2. Product
Essential for e-commerce and catalogs. Marks up name, price, currency, availability, SKU, and ratings (aggregateRating). This is the type that enables product cards with stars and pricing directly in the SERP — one of the rich formats with the highest CTR impact.
3. Article
For editorial content, blogs, and news. Marks up headline, author, publication date, featured image, and publisher. Helps Google understand authorship and content freshness, two signals relevant to E-E-A-T.
4. BreadcrumbList
Describes the navigation path (breadcrumbs). Generates visible breadcrumbs in the search result instead of a raw URL, improving understanding of the site structure and the snippet's appearance. It has one of the best effort-to-benefit ratios of any markup type.
5. FAQPage
Marks up question-and-answer pairs. Its situation changed in 2026 (covered in detail below), but it remains a valid and useful type for semantic understanding and AI engines, even though it no longer generates the expandable rich result in the main SERP.
Practical rule: only mark up what is true and visible on the page. Google penalizes markup for content the user cannot see or that doesn't match what the page actually shows. Structured data must reflect the content — never fabricate it.
How to implement JSON-LD step by step
Implementing JSON-LD is simpler than it looks. The pattern is always the same: a <script type="application/ld+json"> block in the <head> or at the end of the <body>, containing an object that declares its context (@context) and its type (@type).
Step 1. Choose the right type. Identify what the page represents (an article, a product, the organization) and choose a single primary type per entity.
Step 2. Check the properties on Schema.org. Every type has required and recommended properties. Google documents which ones are necessary to be eligible for each rich feature.
Step 3. Write the JSON-LD block. Here's a real, complete example for a blog article:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "Structured Data & Schema.org: JSON-LD SEO Guide",
"image": "https://www.example.com/images/hero.png",
"datePublished": "2026-06-02",
"dateModified": "2026-06-02",
"author": {
"@type": "Person",
"name": "Alfons Marques"
},
"publisher": {
"@type": "Organization",
"name": "Technova Partners",
"logo": {
"@type": "ImageObject",
"url": "https://www.example.com/images/logo.png"
}
}
}
</script>
Step 4. Use absolute URLs for images and links. Properties such as image or url must use absolute URLs, not relative ones, so Google can resolve them correctly.
Step 5. Combine multiple types where applicable. A single page can have both an Article and a BreadcrumbList. You can include multiple <script> blocks or use an array. Keep them consistent with each other.
Step 6. Automate generation. In modern frameworks (Next.js, for example), the standard approach is to generate the JSON-LD from a function that receives the page's data and serializes the object. This avoids manual errors and ensures the markup always reflects the real content.
Implementation checklist
- One clear primary type per entity on the page.
- All required properties for that type are present.
- Absolute URLs in
image,url, andlogo. - The markup matches the visible content.
- Block placed in
<head>or at the end of<body>. - No fabricated data or content hidden from users.
- Validated before publishing (see next section).
How to validate your structured data: Rich Results Test vs. Schema Markup Validator
Never publish structured data without validating it. There are two main tools, and although they are sometimes used interchangeably, they do different things.
| Tool | Maintained by | What it checks | When to use it |
|---|---|---|---|
| Rich Results Test | Only elements eligible for specific Google rich features | To find out if you qualify for a specific rich result | |
| Schema Markup Validator | Schema.org (Google/Microsoft/Yahoo collaboration) | All markup against the full Schema.org specification | To validate general markup correctness |
The distinction matters. Google's Rich Results Test only checks the elements that can generate a rich feature in Google; if you mark up a valid Schema.org type that Google doesn't use for rich results, this tool may say nothing useful or warn you that "no elements were detected." That doesn't mean your markup is wrong.
The Schema Markup Validator from Schema.org, on the other hand, validates all markup against the complete specification, regardless of whether Google uses it to enrich results. It is the right tool for confirming that your JSON-LD is syntactically and semantically valid.
The practical recommendation: use both. Start with the Schema Markup Validator to confirm the markup is correct, then use the Rich Results Test to verify whether it makes you eligible for a specific rich feature. Complete the loop with Google Search Console's Enhancements report, which shows errors and warnings in production for already-crawled pages.
Does structured data still deliver value after the FAQ rich results removal?
This is the question many teams are asking in 2026, and the short answer is yes, absolutely — but it's worth understanding exactly what changed.
The rollback of FAQ rich results wasn't sudden. As early as August 2023, according to the Google Search Central blog, Google restricted FAQ rich results to authoritative government and health websites, and deprecated How-To rich results on desktop. It was a clear signal of where things were heading.
The final blow came in 2026. According to Search Engine Land, on May 7, 2026, Google stopped showing FAQ rich results for all sites. From that point, the retirement timeline is staggered:
- June 2026: Google removes the FAQ report and FAQ support in the Rich Results Test.
- August 2026: FAQ support is removed from the Search Console API.
Now — and this is the nuance many overlook — the Schema.org FAQPage type remains valid, and the markup can stay on your page without causing any issues. What disappears is the expandable visual format in the SERP, not the validity of the markup itself.
What to do with your structured data now
- Don't rush to remove your
FAQPagemarkup. It's valid, it doesn't penalize, and it continues to provide semantic context that AI engines consume to construct and cite answers. - Redirect your effort to types that still generate rich results.
Product,BreadcrumbList,Article, andOrganizationare all alive and delivering returns. - Think GEO, not just SERP. The value of structured data has shifted from "earning a nice dropdown" to "being machine-readable for AI assistants." That trend is accelerating, not reversing.
- Maintain validation discipline. With types entering and leaving rich result support, reviewing Search Console periodically stops being optional.
The strategic takeaway: specific rich formats come and go at Google's discretion, but structured data as your site's semantic layer is an investment that only gains relevance in an AI-assisted search world. Organizations that build clean, accurate, well-validated markup are prepared for whatever comes next.
Conclusion: semantic markup is infrastructure, not a trick
Structured data stopped being a "quick win for gold stars" a long time ago. It's the layer that connects your content to search engines and, increasingly, to generative engines. JSON-LD is the format to use, five types cover most of the value, and validating before publishing is non-negotiable. The removal of FAQ rich results in 2026 is not the end of anything: it's confirmation that what matters is the robustness of the markup, not whatever visual treatment Google happens to display at any given moment.
At Technova Partners we implement and audit structured data as part of technical SEO projects already generating qualified traffic for B2B companies. If you'd like to review the state of your markup or build it right from the ground up, let's talk about your project and we'll tell you where to start.
