JSON-LD

JSON-LD, short for JavaScript Object Notation for Linked Data, is a way of putting structured information about a page into the code, separate from the visible text. It sits in a separate script block, usually in the head of the page, and describes in a fixed format what's on the page: a business, a product, an article, an event or a set of frequently asked questions.

JSON-LD is the format Google recommends for structured data. The vocabulary you use in it comes from Schema.org, a shared standard of the major search engines. So JSON-LD is the packaging, and Schema.org the language.

Why choose JSON-LD?

There are other formats too, such as Microdata and RDFa, which you weave directly into the HTML elements. JSON-LD has the advantage of being separate from your markup. You can add or change it without touching the visible page, which makes maintenance easier. In platforms such as Webflow or WordPress, you can often set it up per page type with dynamic fields, so every page automatically gets the right data. For an online store with hundreds of products, that saves a lot of manual work, because price and stock come straight from the product data.

What does it look like in practice?

A bakery in Edinburgh wants to tell Google and other systems clearly who it is. In JSON-LD, it describes itself as the type Bakery, with:

• its official name, address and phone number, matching its NAP details elsewhere on the web;
• its opening hours for each day;
• the URL of its logo and a few photos;
• links to its social media profiles via the sameAs property.

On its product pages for made-to-order cakes, it can also use Product markup with price and availability. Important: everything in the JSON-LD must also be genuinely visible or correct on the page itself.

JSON-LD and AI search

Structured data helps search engines recognise entities and their properties unambiguously. That fits with how a knowledge graph works: as a network of entities and their relationships. To what extent AI assistants such as ChatGPT, Gemini or Perplexity use JSON-LD directly when putting together answers isn't publicly documented. So it's no miracle cure for AI mentions, but it is a way to present your data as unambiguously as possible to every system that reads it.

Common mistakes and how to avoid them

1. Markup that doesn't match the page: reviews or prices in the code that visitors don't get to see go against Google's guidelines.
2. Syntax errors: one missing comma can make the whole block unreadable. Test with Google's Rich Results Test or the Schema.org validator.
3. Outdated data: opening hours that have been changed on the site but not in the JSON-LD.
4. Expecting markup to always produce an eye-catching search result: Google decides for itself whether to show rich results.

A technical SEO analysis that checks your structured data tracks down missing, incorrect or conflicting JSON-LD, so you know where to step in first.