How to Add Structured Data

Tomohiro Iida · Published June 26, 2026 · Updated June 26, 2026

This guide covers the role of Schema/JSON-LD structured data, how to write it, and how to validate it, from a practical BtoB perspective.

What structured data is

Structured data is a way of labeling the meaning of the information on a web page in a format that search engines and AI can understand. A human reading the text 'Example Co., Ltd.' recognizes it as a company name, but to a machine it is just a string of characters. Structured data attaches labels — 'this is a company name,' 'this is an address,' 'this is a publication date' — using a shared vocabulary. That shared vocabulary is Schema.org.

Schema.org is a vocabulary standard jointly established by Google, Microsoft, Yahoo, and Yandex, defining 'types' such as Organization and Article, and 'properties' such as name and url. Structured data annotates page content using this Schema.org vocabulary.

The key point is that structured data is only a way of conveying information — it does not improve the content itself. Adding structured data to a thin page does not make the content any richer. Structured data's role is to take information already on the page and hand it to machines in a form they won't misread.

It's a common question when structured data is added but ranking doesn't move and no rich result appears — this is expected once you understand what structured data actually does. It's an investment in conveying information accurately, not a guarantee of results.

Types and what each is for

Schema.org defines hundreds of types, but a BtoB site only needs to know a handful in practice. Here are the main ones and their roles.

Organization

Conveys the company name, logo, address, contact details, and official social accounts. This is the core type representing the company as a whole, typically placed on the homepage or the About page. Alongside name, url, and logo, adding sameAs (links to official social accounts or external profiles) can help establish the company as a recognized entity.

WebSite

Conveys basic information about the site as a whole, such as its name and URL. It was once used with SearchAction to produce a sitelinks search box in Google results, but Google has discontinued that feature, so it no longer serves that purpose.

BreadcrumbList

Conveys where a page sits in the site's hierarchy, making a structure like 'Home > Services > SIGNAL' machine-readable. Search-result URL displays sometimes take a breadcrumb format, so this is a type with a clear, easy-to-see benefit relative to its low implementation effort, making it a high priority.

Article / BlogPosting

Conveys a blog or column article's headline, author, publication date, and update date. NewsArticle is a reasonable choice for news-like content, BlogPosting for general blog posts. Indicating an author and publisher helps signal where the information comes from.

FAQPage

Structures the questions and answers on a page. Note carefully: FAQPage is not required, and adding it does not guarantee rich display or AI citation. Google has also discontinued FAQ rich-result support, so adding FAQPage will not produce a rich FAQ display in Google Search. Write the FAQ clearly in the body text first — structured data is only a supplement.

Product / Service

Product conveys a physical product or packaged good; Service conveys a service you provide. For BtoB, consider Product when there is a clearly defined product or plan, and Service for something like consulting or contracted work. Only mark up properties such as price or reviews if they are actually displayed on the page — marking up a price or review that isn't shown is a policy violation and counterproductive.

LocalBusiness

Used by businesses with a physical location or premises that receive visitors, conveying business hours, address, and phone number. BtoB companies may consider this where office visits or local recognition matter; for online-only businesses without visitors, Organization is often sufficient, and there's no need to force in LocalBusiness.

The principle for choosing a type

Choose the type that matches the page's actual nature. Mismatches — putting Product on an article page, or Article on a service page — are a common source of confusion. As a rule of thumb, give each page one primary type, supplemented by BreadcrumbList and Organization where relevant.

The basics of writing JSON-LD

Structured data can be written in JSON-LD, Microdata, or RDFa, but JSON-LD is the common approach today and the one Google recommends. Because it sits in a script tag separate from the body HTML, it is easier to maintain.

A minimal Organization block, placed inside the HTML head (it also works in the body, but head is standard), is a JSON object with three parts: "@context" set to https://schema.org, "@type" set to "Organization," and properties such as name, url, and logo set to the company's actual name, URL, and logo image address.

• @context specifies the vocabulary source, always https://schema.org.
• @type specifies what kind of data this is.
• Everything after that is a set of properties and values defined for that type.

Values must always match what's actually displayed on the page — this is the fundamental rule of structured data.

Organization example

A typical Organization block adds sameAs entries — links to the company's official X (Twitter) and LinkedIn profiles, for example — alongside name, url, and logo.

BreadcrumbList example

A BreadcrumbList lists each level of the site hierarchy as a ListItem, with position for the order, name for the display label, and item for that page's URL. Keep the order and labels consistent with what your breadcrumb navigation actually shows.

BlogPosting example

A BlogPosting block sets headline to match the page title, and datePublished/dateModified to the actual publication and update dates, in ISO 8601 format (for example, 2026-06-26), along with author and publisher information.

Placing multiple types on one page — putting BlogPosting, BreadcrumbList, and, where relevant, FAQPage in separate script blocks, or combining them into one script using @graph — is fine either way, as long as each block's content matches the page's actual content.

What to prioritize for a BtoB site

Not every type needs to be added. For a BtoB site, prioritize the types with clear benefit and low risk of drifting from reality.

Getting Organization, WebSite, and BreadcrumbList right, and adding Article-type markup to articles, is the starting point for most BtoB sites. Mark up Service/Product only to the extent the information is actually shown on the page. Treat FAQPage as a supplement that assumes the body copy is already solid, not as something that pays off just by adding it.

Common implementation mistakes and how to validate

Here are the failure patterns that come up most often in structured-data reviews. Most trace back to a mismatch with reality or a skipped validation step.

Common mistakes

• Marking up information not shown on the page — writing a price, rating, or review into structured data alone that isn't displayed in the body text. This is a policy violation and the mistake to avoid above all others.
• Missing required properties — leaving out properties recommended or required for a given type, so it isn't recognized as eligible.
• Syntax errors — broken JSON from mismatched commas, unclosed quotes, or mismatched braces.
• Type/reality mismatch — attaching a type that doesn't fit the page, such as Product on an article page or Article on a service page.
• Inconsistent URLs or dates — an item URL that doesn't exist, or a datePublished that doesn't match the actual publication date.
• Markup on a noindex page — adding structured data to a page that isn't indexed in the first place won't make it eligible for display.
• Missing absolute image paths — writing a logo or image as a relative path that search engines can't resolve.

How to validate

Always check with the following tools before and after publishing.

• Rich Results Test (Google) — enter a URL or code and it detects structured data Google can recognize for rich results, showing errors and warnings. Use it to check types tied to rich results.
• Schema Markup Validator — checks syntax and type validity against the Schema.org spec, including types not eligible for rich results.
• Search Console's Enhancements report — after publishing, continuously check the errors and warnings Google actually detected, with status shown per detected type, such as Breadcrumbs.

Validation isn't a one-time step. Site changes or template updates can break structured data, so keep checking Search Console periodically after publishing too. Not treating it as 'done once it's added' is the key to keeping structured data healthy.

Structuring for AI

Questions like 'does adding structured data get us into AI search' come up more often in the context of AI search (AI Overviews, ChatGPT, Perplexity, and similar). This is an area with a lot of misunderstanding, so it's worth spelling out carefully.

First, as a starting premise: there is no guarantee that adding structured data gets a page cited by AI. What AI chooses to cite depends on each service's own algorithm, not just on whether structured data is present.

That said, structured data can still help convey facts to AI without misinterpretation. When facts such as company name, services, location, contact details, and publication date are organized as machine-readable text, that gives AI a clue for interpreting the page's content. Conversely, when key information exists only as text inside an image or a PDF, it becomes hard to read even before structured data enters the picture.

What matters for conveying facts to AI is not just the 'format' of structured data but the organization of the body text itself. Writing that helps both machines and people includes:

• Leading with the conclusion (answer-first)
• Covering one point per paragraph
• Including specific numbers, conditions, and sources
• Answering a question-form heading concisely right after it

Structured data's role is to attach the correct label to facts that are already organized this way. Body-text organization (the substance) and structured data (how it's conveyed) work together — adding structured data alone will not substantially change how much gets through to AI.

Priority order for fixing things

Trying to fix everything about your structured data at once tends to balloon the workload. Splitting it into stages by impact and effort works better.

The order matters. Starting with individual types like FAQPage or Product before the foundation (Organization, BreadcrumbList) is solid makes it hard to see the effect of anything. For every type, confirm it matches reality and isn't broken, using the Rich Results Test and Search Console, before moving to the next stage.

Structured-data work pays off more when it's done alongside body-copy improvements (making the substance clearer). Structured data is an improvement in how you convey things; pair it with improving what you're conveying.

Summary

• Structured data (Schema/JSON-LD) is a supplement that helps convey page content accurately to search engines and AI. It does not guarantee ranking, rich results, or AI citation.
• Schema that doesn't match reality is counterproductive. Only mark up information actually shown on the page.
• JSON-LD is the common format, placed inside the HTML head.
• For BtoB, Organization, WebSite, and BreadcrumbList are the high-priority basics. Add Article/BlogPosting to articles. FAQPage is neither required nor a guarantee — get the body copy right first.
• Validate with the Rich Results Test and the Schema Markup Validator before publishing, and keep checking via Search Console after.
• For AI, structured data (how you convey it) and body-copy organization (what you convey) both matter.
• Work through it in order — foundation, then articles/services, then validation and operation — alongside body-copy improvements.

Frequently asked questions

Does adding structured data raise search ranking?

Not necessarily. Structured data helps convey content accurately; it does not directly raise ranking. Ranking is determined by other factors, such as content quality and match with search intent.

Does adding structured data guarantee a rich result will display?

No. Whether a rich result displays is the search engine's own decision, and it may not display even when the structured data is correct. Being recognized as eligible by the Rich Results Test and actually being displayed are two different things.

Should I write JSON-LD or Microdata?

JSON-LD is the common approach today, and Google recommends it. Because it can be written separately from the body HTML, gathered inside the head, it's easier to maintain.

Should I add FAQPage structured data?

It's not required. Adding it does not guarantee a rich display or AI citation. Write your FAQ clearly as body copy first, and treat structured data as a supplement.

Can I mark up a price or review that isn't shown on the page?

No. Writing information into structured data that isn't shown on the page is a policy violation, and it works against you in evaluation too. Only mark up information that's actually displayed.

How do I check whether structured data is correct?

Before publishing, check with Google's Rich Results Test and a schema validation tool (Schema Markup Validator). After publishing, keep an eye on the Enhancements report in Search Console for detected errors and warnings.

What structured data should a BtoB site add first?

Organization (company information), WebSite (site information), and BreadcrumbList (breadcrumbs) are the high-priority basic set. Add Article/BlogPosting to article pages. Mark up Service/Product carefully, only to the extent it matches reality.

Does adding structured data get a page cited by AI?

There's no guarantee. What AI chooses to cite is up to each service. That said, when facts are organized in machine-readable form, it can give AI a clue for interpreting the page. Structured data (how you convey it) and body-copy organization (what you convey) both matter.