Schema (structured data) markup: rich results and AI visibility

Picture a two-person workshop that makes handmade leather bags. The products are lovely and the pages are tidy, yet when someone searches on Google the result is just a blue headline and two lines of text. Right below it sits a much larger competitor showing star ratings, stock status, a breadcrumb trail and a thumbnail image. Same query, completely different visibility. The gap usually isn't a better product or a bigger budget; it's whether the page explains itself clearly to machines. That is exactly what Schema (structured data) markup does: it tells search engines and AI, in a precise language, what your content actually means.

This guide walks through what structured data is, how it feeds rich results and AI visibility, which Schema types an SME should prioritise, why JSON-LD is Google's recommended format, how to add markup to your site and test it, and which mistakes can land you in trouble. No invented statistics or inflated promises, just the mechanism, example types and a practical starter plan.

1Content2JSON-LD schema3Engines understand4Rich results & AI

What is structured data (Schema)?

Structured data means labelling the information on a page with a standard vocabulary that machines can understand. That vocabulary is Schema.org, a shared dictionary built jointly by Google, Microsoft, Yahoo and Yandex. It defines entities such as "this is a product", "this is a business", "this is an event", along with their properties: name, address, rating, date, availability. When a person reads your text they can tell that one line is a price and another is an opening hour; a search engine can't always extract that reliably from raw HTML. Schema closes that gap.

Put simply: ordinary text shows search engines what you wrote; structured data tells them what you mean. On a restaurant page the phrase "Closed on Mondays" becomes, when marked up correctly, not a sentence but a machine-readable "opening hours" value. That is why structured data has become an increasingly central part of technical SEO.

How a search result looks today can change tomorrow, but stating clearly what your content means always pays off.

Rich results: becoming more visible in search

The most visible payoff of structured data is rich results: eye-catching search appearances that go beyond the plain blue link. The best-known examples are star ratings, product cards with stock and price, breadcrumb trails, event dates and recipe cards. They take up more space on the results page, can attract more clicks, and are a concrete return on your search engine optimisation effort.

One important caveat, though: the catalogue of rich results is not fixed; it changes over time and Google retires some appearances. The once-popular FAQ rich results, for instance, were first limited to well-known, authoritative government and health sites, and then removed from search entirely as of 7 May 2026; the "HowTo" appearance had already been deprecated earlier. Your markup does no harm, but it no longer produces that visual card. The lesson is clear: rather than chasing a specific visual flourish, describe the meaning of your content accurately, because that is where the durable value lives.

AI visibility: helping generative engines understand you

The second and fast-growing benefit of structured data is on the AI side. Generative engines like ChatGPT and Gemini, and Google's AI Overviews summary, have to quickly work out which business does what and which product solves which problem when they compose an answer. A clean entity definition, who you are, what you sell and where you are, makes it easier for these engines to understand you correctly and cite you.

This overlaps directly with generative engine optimisation (GEO): where classic SEO aims to appear in the ranking, GEO aims to appear in the AI's answer. Structured data is the common ground that serves both. When a language model reads your page, a well-marked-up Organization or Product entity is far easier to parse than a loose block of text. Rich results may come and go, but the need for machines to understand your content only grows.

The most useful Schema types for SMEs

Schema.org has hundreds of types, but an SME doesn't need to use them all. Depending on your business, a handful do most of the work:

• Organization and LocalBusiness: your business name, logo, address, phone and opening hours; the most basic markup for local businesses. Together with your Google Business Profile and local SEO work, it builds a strong identity.
• Product and Offer: for e-commerce, the product name, brand, availability and price fields. You populate the price field from your own current catalogue; the point is to define the field correctly rather than invent a number.
• Review and AggregateRating: ratings drawn from genuine customer reviews. Mark up only real reviews that are actually visible on your site.
• Article, BreadcrumbList and Event: the most appropriate types for blog posts, on-site navigation structure and events respectively.
• FAQPage: even though it no longer produces a rich result, it still communicates the question-and-answer structure of your content to machines.

JSON-LD: Google's recommended format

Structured data can be added to a page in three formats: JSON-LD, Microdata and RDFa. Google explicitly recommends JSON-LD. The reason is practical: instead of scattering markup through your visible HTML, JSON-LD gathers it into a single <script type='application/ld+json'> block placed on the page. That keeps the data decoupled from the design, so your markup doesn't break when you change a template, and it is far easier to manage from one place.

Microdata and RDFa are still valid, but because they embed attributes directly among your HTML elements they are harder to maintain. If you are starting fresh, beginning with JSON-LD is almost always the right choice.

How to add Schema to your site

You don't have to write code. Depending on your team's technical capacity, there are three routes.

Hand-coded JSON-LD

Hand-written JSON-LD gives you the most control. You choose the right type, fill in the fields and add the script block to the relevant page. It is ideal for a small number of critical pages, such as Organization on the homepage and LocalBusiness on the contact page.

CMS and plugins

In WordPress, Yoast or Rank Math; in Shopify, built-in templates or dedicated apps generate most markup automatically. This is usually the fastest way to add product, article and breadcrumb data without touching code by hand.

Google Tag Manager

If you can't edit the site's code directly, you can inject JSON-LD through a custom HTML tag in GTM. It is flexible, but you must carefully check that the tag fires on the right page and that the data stays consistent with the page.

Testing, validation and monitoring

Once you've added markup, always validate it; a small syntax error can invalidate the whole block. Three tools are used together:

• Rich Results Test: Google's tool; it shows which rich result types your page is eligible for.
• Schema Markup Validator: checks general Schema.org syntax and catches errors that aren't Google-specific.
• Google Search Console: after you go live, the "Enhancements" reports show valid and invalid items on real user data. Monitoring Search Console regularly lets you spot a problem before it grows.

Common mistakes, Google's policy and a starter plan

Structured data is a powerful tool, but it backfires when abused. Google's structured data policies are clear, and a violation can trigger a manual action (a penalty). The most common mistakes are:

• Marking up what isn't on the page: only mark up content the user actually sees on the page; marking up hidden or different information is a policy violation.
• Fake or self-serving ratings: invented reviews and inflated AggregateRating values count as spam.
• Irrelevant or spammy markup: adding random types to every page creates risk, not benefit.

A healthy starter plan looks like this: first add Organization to the homepage and LocalBusiness to your local branch; then Product/Offer on your most important product or service pages, and Review if you have genuine reviews; then bring in breadcrumbs and Article. Run each step through the Rich Results Test and watch it in Search Console. Starting small and growing through validation is far more robust than adding everything at once and chasing errors later.

Frequently asked questions

In short, Schema (structured data) markup is not a magic ranking trick; it is the standard way to tell search engines and AI clearly what your content means. The look of rich results may change over time, but the need for machines to understand you correctly is here to stay, which is why marking up meaning always pays off. Start with small, validated steps. As you capture the demand this visibility brings, a CRM and marketing hub like Rocketly completes the loop by making the interest arriving from search and AI channels visible in one unified inbox.