Structured data, also called schema or schema markup, is a label for your website written in a language machines read fluently. Your visitors never see it. It sits in the code of the page and tells Google, Bing and AI assistants, in no uncertain terms: this is a business, here is its name, here is where it is, here is what it does, here is what it charges.

Picture the label on a tin. A human can often guess soup from the picture. The barcode is there so the till doesn't have to guess. Schema is the barcode for your website, and for a small business it is one of the cheapest, most durable bits of SEO work you can do. Here is what it is, the four types worth having, and how to check what your site already has.

What is structured data, exactly?

Under the surface, it is a small block of code, usually a format called JSON-LD, placed in the head of a page. It uses a shared vocabulary from schema.org, a dictionary of types that Google, Microsoft and others agreed on years ago, so every search engine and AI system understands the same labels.

Three things worth knowing before we go further:

• It is invisible to visitors. Nothing about your design changes.
• It describes what is already on the page. It is a translation, not a place to make claims your page doesn't back up.
• It is a one-off job that quietly pays out for years, as long as someone keeps it accurate when your details change.

Why should a small business care?

Two reasons, and the second one is newer than the first.

In Google, schema makes you eligible for rich results: the FAQ dropdowns, review stars, business details and article panels that make some listings visibly bigger than others. Google has been clear that schema by itself is not a ranking factor, but richer listings earn more clicks from the same position, and clear machine-readable facts reduce the chance Google misunderstands what you do.

In AI answers, the case is simpler. When ChatGPT, Perplexity or Google's AI Mode assembles an answer about your trade in your town, it works from what it can read and verify. A page that states its facts in schema is harder to misquote than one that buries them in brochure copy. This is a large part of what we do under the banner of SEO and GEO: making businesses legible to machines without making the website any less human.

The four schema types that matter for a small business

Schema.org lists hundreds of types. You can ignore nearly all of them. For a typical small business, four do the heavy lifting.

1. Organization (or LocalBusiness): who you are

This is your business card in code: name, logo, address, phone, opening hours, the area you serve. If you serve customers at or from a physical location, use LocalBusiness (or a specific flavour of it, like Dentist or Plumber); otherwise plain Organization is fine. It supports your Google Business Profile, feeds knowledge panels, and gives AI assistants a single trustworthy source for your basic facts. Every business should have this one, sitewide, usually via the template.

2. Service: what you do

Service schema describes each thing you offer: what it is called, who provides it, where, and (if you are brave enough to publish it) the price or price range. It helps search engines connect "emergency boiler repair Taunton" to your actual boiler repair page rather than your homepage, and it gives AI answers concrete services to name instead of a vague "they do heating stuff".

3. FAQPage: the questions you answer

If a page genuinely answers questions (a pricing page, say, or a service page with a Q&A section), FAQPage schema marks each question and answer pair. In Google this can show your FAQs directly in the results; in AI answers, question-shaped content is exactly the shape assistants are looking for. The rule: the questions and answers must appear on the page, word for word. No marking up FAQs that exist only in the code.

4. Article: the things you publish

Every blog post should carry Article (or BlogPosting) schema: headline, description, publish date, author, publisher. It tells machines the content is editorial, when it was written, and who stands behind it, which matters for freshness and trust, in Google and AI systems alike. If you publish regularly, your CMS should be doing this for you automatically; ours does on every site we build.

What does schema actually look like?

Here is a short, realistic example for a local business. This is the whole trick. There is no more magic beneath it:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "LocalBusiness",
  "name": "Westfield Plumbing & Heating",
  "url": "https://www.example.co.uk/",
  "telephone": "+44 1823 000000",
  "priceRange": "££",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "12 High Street",
    "addressLocality": "Taunton",
    "addressRegion": "Somerset",
    "postalCode": "TA1 1AA",
    "addressCountry": "GB"
  },
  "areaServed": "Somerset",
  "openingHours": "Mo-Fr 08:00-17:00"
}
</script>

Readable, isn't it? Every line is a plain fact about the business, stated once, in a format no machine can misread. That is all schema is.

How to check what your site already has

Before adding anything, find out where you stand. Two free tools, five minutes:

1. validator.schema.org: paste in any page URL and it lists every piece of structured data it finds, with errors flagged. This shows you the raw truth.
2. Google's Rich Results Test (search.google.com/test/rich-results): the same idea, but focused on which rich results your page is eligible for in Google specifically.

Run your homepage, one service page and one blog post through both. Three common outcomes: nothing at all (common, easily fixed), auto-generated schema from your platform (check the details are actually right; we regularly find sites confidently declaring the wrong phone number), or duplicate conflicting schema from stacked plugins (worse than nothing; pick one source of truth).

Schema also pairs naturally with the rest of the machine-readability work: server-rendered content, llms.txt, honest sitemaps. We covered the full list in our AI SEO checklist.

Common questions

Will structured data improve my Google rankings?

Not directly. Google has said schema is not a ranking factor on its own. What it does is make your listing eligible for richer results (stars, FAQs, business details), which tend to earn more clicks than plain blue links, and it removes ambiguity about who you are and what you do. Better understanding plus better click-through is worth having, even without a rankings boost.

Can I add schema myself without a developer?

Sometimes. Many website platforms and plugins generate basic Organization and Article schema automatically, and there are free JSON-LD generators for simple cases. Where it gets fiddly is placing the code correctly, keeping it accurate as your business changes, and avoiding conflicting or duplicate schema from multiple plugins. If your platform handles it, check the output; if not, it is a small, well-defined job for a developer.

Does schema have to match what is on the page?

Yes. Schema must describe content that is actually visible on the page. Marking up five-star reviews you do not display, or FAQs that appear nowhere in the copy, goes against Google's guidelines and can earn a manual penalty. Schema is a translation of the page, never an embellishment of it.

Does structured data help with ChatGPT and AI search?

It helps. AI systems read the same web pages Google does, and unambiguous machine-readable facts (your name, location, services, prices) are easier to quote correctly than facts buried in marketing copy. Schema is not the whole job of being visible in AI answers, but it is one of the cheapest ways to make sure that when you are mentioned, the details are right.

Want your site labelled properly?

We are Bare Creative: Jack and Carly, a husband-and-wife web studio in Somerset. Schema is built into every site we ship, and our care plans keep it accurate as your business changes, because a wrong phone number in code is worse than no code at all. If you would like us to check what your site is currently telling the machines, drop us a line and we will send you an honest, plain-English rundown.