Structured Data SEO: Help Google Understand Your Pages

After 13 years of auditing sites, the same three markup faults still turn up on almost every small business site I open. Nothing is marked up at all. Something is marked up but says the opposite of what the page says. Or a plugin wrote the markup two years ago and nobody has read it since.

None of those is a disaster, and all three are an afternoon of work to fix. What they share is a misunderstanding of what the markup is for. So before the detail, the plain answer: structured data is a labelling system, not a ranking switch, and Google guarantees you nothing for adding it.

What follows is written for an owner or an in house marketer who has been handed a report that says add schema and wants to know what that actually changes.

What Structured Data Actually Does

Structured data is a standardized format for providing information about a page and classifying the page content. That is Google's own wording in its introduction to structured data markup, and the second half is the part most reports skip. Classifying. Not ranking.

Structured data, schema markup and vocabulary

Three terms get used interchangeably and they do not mean the same thing. Structured data is the markup sitting in your page. Schema markup is structured data written in the Schema.org vocabulary. The vocabulary is a shared dictionary of types and properties, and it is not owned by Google. Schema.org is a collaborative, community activity with a mission to create, maintain, and promote schemas for structured data on the Internet. That is the project's own summary of itself on its about page.

Labels for machines, not keywords for ranking

Think of the markup as labels stuck on parts of a page: this block is the business name, this number is the price, this person wrote it, this paragraph is the answer to a question. Google still reads your page. The labels stop it guessing, and they make certain search features possible in the first place.

Why Google Needs Labels, not Guesswork

A page is a wall of text to a machine until something tells it what each part is. Take a local plumber. The page mentions a business name, an address, a phone number, a list of services and a review score. A person separates those things by reading the layout. A machine has to infer them from position, formatting and wording, and inference is where mistakes happen. Two businesses with similar names in the same suburb are the classic case, and a misread address puts a business in the wrong map area.

Labels also decide whether a feature can exist at all. Rich results such as review stars, event listings, recipe times and FAQ blocks are built from markup. No markup, no eligibility, however well the page reads. That is the honest case for doing this work, and it is a case about how your listing appears rather than where it ranks.

The Three Formats Google Supports

Google supports three ways of writing the same facts, and only one of them is worth learning in 2026.

JSON-LD, microdata and RDFa

Microdata and RDFa sit inside your HTML attributes, wrapped around the visible elements they describe, which makes them harder to read and easier to break. JSON-LD is a script block, usually in the head or the footer, written as plain text you can copy, compare and keep in version control.

Why JSON-LD is the practical default

For a business site the question is not which format is theoretically tidier, it is which one survives a redesign. A script block in one place is easier to review, easier to hand to a developer, and far less likely to be stripped out when someone rebuilds a template. I write JSON-LD by default and treat the other two as legacy.

Schema Types Worth Adding First on a Service Site

There are hundreds of types. You need a handful, and you need them complete.

Organisation, LocalBusiness and Person

Start with the entities that describe who you are and where you operate. Organisation markup carries your name, logo, website and contact details. LocalBusiness narrows that to a physical location with an address, a service area and opening hours, which is the markup that supports local search features. Person describes the human behind the business, and on a consultancy site it links the advice to a named practitioner rather than to a faceless entity.

Article, FAQ and breadcrumb markup

Then the content layer. Article marks up editorial pages so a machine knows the headline, the author and the publication date. FAQ marks up question and answer pairs that already exist on the page. Breadcrumb markup describes the path a reader followed, which is why it can replace a raw URL line in the results. If your page carries a FAQ section, that is a section waiting for markup.

Product and Review markup for ecommerce

On a store, Product markup carries the name, price, availability and identifiers, and it is what makes a shopping result informative rather than a blue link. Review and rating markup has stricter rules than most people expect, and it belongs on pages where reviews of that specific product are visible. Rating markup on a service page with no visible reviews is the fastest way to lose eligibility.

How to Add Structured Data to a Real Site

Plugin output versus hand written JSON-LD

A plugin is a reasonable starting point and a poor finishing point. Most SEO plugins emit a default set for organisation, website and breadcrumbs, and the defaults are frequently wrong: a placeholder logo, a founding date nobody has checked, an address from a previous office. Hand written markup for the two or three types that matter usually beats twenty automatic types nobody has read. Either way, read what is on the page.

Where the markup lives and how it nests

The script block sits in the page source, and types can reference each other: a Product can name an Organisation as its seller, and an Article can name a Person as its author. Referencing is where plugins quietly go wrong, because a reference to an entity that does not exist on the site produces a dangling link. Keep the number of types small enough that you can name every property inside them.

The Rules That Decide Whether You Get a Rich Result

Eligibility is stricter than most people expect, and the rules are published in Google's general structured data guidelines.

Markup must match visible content

Don't mark up content that is not visible to readers of the page. That is the rule in Google's own guidelines, and it is why invented review scores and hidden price markup are treated as spam rather than as clever optimisation. If a fact is worth marking up, it should be worth printing on the page.

Required properties and eligibility

Items that are missing required properties are not eligible for rich results. Every type carries its own required set, and a half finished type is worse than no type at all, because it consumes the page's markup and returns nothing. Add what you can complete properly.

Quality guidelines and manual actions

Beyond the required properties sit the general quality rules: no misleading markup, no fake reviews, no marking up other people's content as your own. Breaches can cost you rich result eligibility across the site and, in serious cases, end in a manual action. This is one of the few areas of SEO where the cost of getting it wrong is a real penalty rather than a missed opportunity.

Testing Properly with the Rich Results Test

Test one page at a time with the Rich Results Test, fix the errors, then test the template you copied the markup into. Then, and this is the step people skip, wait. Google does not guarantee that your structured data will show up in search results, even if your page is marked up correctly according to the Rich Results Test. A clean test means your markup is valid and the item is eligible. It is not a promise that anything changes in the results.

Checking that properly is where an SEO audit that checks your markup against the live page earns its place, because the failure I see most often is markup that was valid on a staging site and wrong on the live one.

What Structured Data Will Not Do

It will not lift your rankings on its own. It will not fix thin content, a slow page or a weak title. It will not make a page eligible for a feature it does not qualify for. What it can change is how a result presents, and presentation moves click through rate. Google's documentation on structured data cites Nestlé's own measurement: "Nestlé has measured pages that show as rich results in search have an 82% higher click through rate than non-rich result pages." That is a large figure from one company's data, so I read it as evidence that presentation matters rather than as a forecast for your site. The clicks still have to be worth having, and the page still has to answer the query.

Does Schema Markup Help AI Search Visibility

Indirectly, and less than the people selling AI optimisation packages suggest. Markup helps a machine classify the page, which is a sensible supporting signal for any system that has to decide what a page is about. What it is not is a shortcut. Google's guidance on AI features and your website states that there are no additional requirements to appear in AI Overviews or AI Mode, and no other special set up is needed beyond being indexed and eligible for a snippet. My advice is to treat markup as hygiene: do it properly because it makes your pages machine readable, and do not expect it to be the reason you get mentioned. If AI search visibility is the actual goal, how I approach AI search visibility starts with content and entities rather than markup.

A 30 Day Schema Plan for a Small Site

Week one: inventory what your plugin already emits, and delete the types you cannot complete. Week two: write Organisation, LocalBusiness and Person properly, with the real address, the real logo and the real contact details. Week three: add Article and Breadcrumb to your post template, and FAQ to the two pages that already carry visible questions. Week four: run every template through testing, fix what fails, and write the property list into your site documentation so the next person does not have to guess.

If you want the wider context, the wider technical SEO checklist covers the rest of the plumbing, and the questions I get asked most covers how I work with clients.

Structured Data SEO FAQ

Will structured data improve my rankings?

No, not directly, and anyone promising that is overselling. Google's own documentation frames structured data as a way to classify page content so it can power search features and rich results. The measurable upside is presentation and click through rate, not position. In Google's cited case study, rich result pages saw an 82 percent higher click through rate than non-rich result pages, which matters commercially even when the ranking is unchanged.

Which schema type should I add first?

For a local service business, Organisation and LocalBusiness first, then Person for the consultant or author, then Article on editorial pages and Breadcrumb for navigation. Each one has required properties. If any are missing the item is not eligible, so add types you can complete properly rather than ten half finished ones.

Is JSON-LD better than microdata?

Google supports JSON-LD, microdata and RDFa, and recommends JSON-LD. It keeps the markup in one place, is easier to review and edit, and is far less likely to break your page templates when someone updates a layout.

Why did my markup test clean but no rich result appeared?

Testing confirms syntax and eligibility, not display. Google states plainly that it does not guarantee structured data will show up in search results even when a page passes testing, and a missing required property makes the item ineligible outright.

Can I add markup for content that is not on the page?

No. Google's quality guidelines require markup to be a true representation of the visible page, and marking up content readers cannot see risks losing rich result eligibility or being treated as spam.

Does schema markup help AI search visibility?

It helps machines classify the page, which is a reasonable supporting signal, but Google states there are no additional requirements or special optimisations needed to appear in AI Overviews or AI Mode beyond being indexed and eligible for a snippet. Treat schema as hygiene, not as an AI shortcut.