The Complete Guide to Schema Markup and Structured Data

Most guides to schema markup are quietly out of date, and following them will have you spend a weekend adding code for search features that no longer exist. Google removed FAQ rich results in May 2026 and killed HowTo results back in 2023, yet the advice to add both is still everywhere. So this guide does two things. It explains structured data from the ground up, and it tells you the truth about what actually earns anything in Google today, in 2026, rather than what worked five years ago.

Structured data is one of the most useful tools in on-page SEO, but only when you understand what it does and, just as important, what it does not. Let us clear both up.

What is schema markup and structured data?

Structured data is a standardized way of labeling the content on a page so machines can understand what each part means, rather than just reading it as text. Schema markup is the code you add to create those labels, using a shared vocabulary from schema.org, a project backed by Google, Microsoft, and others. In plain terms, structured data turns "Acme Widget, 4.5 stars, $29" from a sentence a machine has to interpret into explicit fields it simply reads: this is a Product, this is its rating, this is its price.

One distinction causes most of the confusion in this topic, so fix it early. Schema markup is the code on your page. A rich result is the enhanced listing Google may show because of that code, the star ratings, the price, the breadcrumb trail that make a result stand out. The markup is the cause. The rich result is one possible effect, and as we will see, Google grants far fewer of those effects than it used to.

The recommended format is JSON-LD, a block of JavaScript Object Notation placed in a script tag on the page. Two older formats, Microdata and RDFa, embed the labels inline in your HTML, but Google recommends JSON-LD because it keeps the structured data in one clean block separate from your content, which is easier to add, read, and maintain.

Does schema markup improve your Google rankings?

No, not directly, and anyone who tells you otherwise is selling something. Google has been consistent for years that structured data is not a ranking factor. Adding schema will not lift your position in the results on its own. What it can do is change how your result looks, and that is where the real value sits.

The benefit is the rich result. A listing with star ratings, a price, or a recipe image takes up more space and draws more eyes, which can lift your click-through rate even when your ranking position has not moved. More clicks from the same position is a genuine win, and it is the honest case for structured data. Google's documentation on structured data frames it exactly this way: eligibility for search appearance features, not a ranking boost.

There is a hard rule attached. Your structured data must match the content visible on the page. Marking up a rating that does not appear to users, or a price that is not really there, violates Google's policy and can earn a manual action. On the AI side, worth noting because the myth is spreading fast: Google's guidance states that no special schema is required for AI Overviews or AI Mode, though any structured data you use should still match your visible content. Do not add markup hoping to bribe the AI. Add it to describe what is honestly on the page.

Does structured data help with AI Overviews and AI search?

Partly, and precision matters here because the hype has run well ahead of the evidence. Google has stated directly that no special schema markup is required to appear in AI Overviews or AI Mode, so you cannot mark your way into an AI answer. What structured data genuinely does is make your page easier for any machine to parse without guessing, and that unambiguous clarity is a general advantage as search shifts toward generated answers. It also reaches beyond Google. Other engines and the retrieval crawlers behind AI assistants still read structured data, including some types Google has stopped rendering. So the sensible stance is to implement the schema that describes your page honestly, expect no magic from it inside AI results, and treat any AI benefit as a bonus rather than the reason you did it.

Which structured data types still earn rich results in 2026?

This is the section that matters most, because the list has shrunk, and building your strategy on the old one wastes real effort. Google has spent three years pruning its rich result portfolio, so treat any pre-2024 guide with suspicion.

The types that still reliably generate rich results in 2026 are the commercially useful ones: Article, Product, Review and AggregateRating for star ratings, BreadcrumbList, VideoObject, Event, Recipe, and LocalBusiness with Organization. Product schema has become more important, not less, as Google leans into AI shopping surfaces. These are where your implementation effort belongs.

Now the graveyard, so you stop wasting time on it. FAQ rich results were removed on May 7, 2026, ending even the narrow eligibility that government and health sites had kept. HowTo rich results were deprecated in 2023 and appear on no surface today. In June 2025, Google retired a further batch of types including Course Info, Estimated Salary, Claim Review, and Special Announcement. Here is the important nuance, though: FAQPage and HowTo are still valid schema.org types. The markup will not throw errors, will not trigger a penalty, and does not need to be ripped out if you already have it. Other engines and AI crawlers may still parse it. What changed is that Google no longer renders a visible feature from it, so you should simply stop adding it in the hope of a rich result that will never come.

How do you add schema markup to your site?

You add schema markup by generating a JSON-LD block for the correct type, placing it on the page, and validating it. The process is the same whether you hand-write it or use a plugin, and getting the order right saves you from the most common errors.

A minimal JSON-LD block sits inside a script tag and looks like this:

{
  "@context": "https://schema.org",
  "@type": "LocalBusiness",
  "name": "Acme Plumbing",
  "telephone": "+1-312-555-0100",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "12 Main Street",
    "addressLocality": "Chicago",
    "addressRegion": "IL",
    "postalCode": "60601"
  },
  "openingHours": "Mo-Fr 08:00-18:00"
}

Notice that every value in that block is something a visitor can also see on the page: the business name, the phone number, the address, the hours. That is the standard to hold every field to, and it is the single habit that keeps you clear of trouble.

Work through it in order. First, choose the right type from the list that still earns results, matched to what the page actually is. Second, generate the JSON-LD, either by hand from the schema.org reference, with a generator tool, or through your CMS. Third, make every field match the visible content exactly, because this is the rule that protects you from a manual action. Fourth, validate it: paste the URL or code into Google's Rich Results Test to confirm eligibility, and use the schema.org validator to check the syntax. Fifth, monitor the enhancement reports in Google Search Console over the following weeks to catch errors at scale and confirm Google is reading your markup.

Common schema markup mistakes to avoid

Even a valid block of schema can work against you if you get the fundamentals wrong. A few mistakes account for nearly all the trouble.

The worst is marking up content that is not visible on the page, which is the fast route to a structured-data manual action. Close behind is choosing the wrong type, or stuffing a page with every type you can think of rather than the one that fits, which confuses more than it helps. Then there is the 2026 classic: adding FAQ or HowTo schema and expecting a rich result, when neither has produced one for months. Skipping validation is another, since a single malformed field can void the whole block silently. And letting your markup drift out of sync with your content, so the price in your schema no longer matches the price on the page, quietly breaks the trust the markup was meant to build. Structured data is a promise to the search engine about what your page contains. Break the promise and you lose the benefit, or worse.

Structured data is a description, not a trick

Here is the guide in one idea. Schema markup does not raise your rankings, and it never conjured a rich result out of thin air. It describes your page accurately so machines can understand it, and when the type still qualifies, Google rewards that clarity with a richer listing that earns more clicks. In 2026 the winning move is narrower and simpler than the old advice: mark up the commercially real things (products, reviews, articles, local business details) with clean JSON-LD that matches your visible content, validate it, and skip the deprecated types entirely.

Because structured data errors hide across templates and silently disable your rich results, the efficient way to manage it is to crawl the whole site and check every page at once. Crawlpit Monster's on-page SEO audit reports the structured data it finds on each page and flags what is missing or malformed, so you fix the template rather than hunt page by page. It is one of the checks in a full technical SEO audit, runs on your own machine, and works on staging sites.