JSON-LD or Microdata.

Both express the same schema.org vocabulary; the difference is where the markup is written — JSON-LD in a separate block, Microdata embedded in the HTML.

Structured data is the layer that tells a machine what a page is about: is this an article, a business, a product? The vocabulary is shared — schema.org — and the two formats are two ways of writing the same vocabulary. The choice affects maintenance, not meaning.

Why JSON-LD became common

Sitting in a separate block makes it independent of the visual structure. When the design changes, a section moves or a component is rewritten, the markup does not break. Nodes can also be linked to one another with `@id`: organisation, person, site and page become a single graph, and the same entity is referred to with the same identity on every page.

On this site the graph is built exactly that way, and every route carries a WebPage node. The full basics and the concept of an entity are written separately.

What still holds for Microdata

Because Microdata writes the markup onto the content itself, it keeps what is shown and what is declared consistent by nature. When `itemprop="name"` is written on a heading element, the declared name is the name on screen; they cannot diverge. In JSON-LD that consistency has to be maintained by hand.

In practice this points to JSON-LD’s greatest risk: declaring something that is not on the page is technically easy. The rule is clear — the markup describes the page’s content, it does not decorate the page. How errors are found is described separately.

The decision

For a new build, choose JSON-LD: it is Google’s recommended format, its maintenance sits in one place, and it lends itself to building a graph. An existing site carrying Microdata, however, should not be migrated for the format alone — changing working markup has no measurable benefit.

Criterion

1. Where it lives — JSON-LD: In a <script type="application/ld+json"> block · Microdata: Inside the attributes of HTML elements
2. Tie to the markup — JSON-LD: Independent — survives changes to the visual structure · Microdata: Tight — moving a div can break the markup
3. Google’s recommendation — JSON-LD: The recommended format · Microdata: Supported, not recommended
4. Relations between nodes — JSON-LD: Stable links can be built with @id · Microdata: Possible with itemref, but fragile
5. Maintenance — JSON-LD: Gathered in one place · Microdata: Scattered across the markup
6. Consistency with visible content — JSON-LD: Must be kept by hand — it sits elsewhere · Microdata: Natural — it is written on the same element
7. Risk of misuse — JSON-LD: Easy to declare what is not on the page · Microdata: Harder — the markup is tied to the content

FREQUENTLY ASKED QUESTIONS

Which one ranks better?

Neither. Structured data is not a direct ranking factor; its job is to tell a machine what the page is. Both formats carry the same information and are read the same way.

Can I use both?

Technically yes, but it is not advisable: declaring the same entity twice invites contradiction and doubles the maintenance. Choosing one format and staying with it is the right approach.

What is the risk with JSON-LD?

Being detached from the visible content. Because the markup sits in a separate block, it is easy to declare a frequently asked question that is not on the page or a rating that does not exist — and that is a violation search engines apply manual actions for.