Most homepages carry no structured data: 9 of 14

FIELD TEST · 2026-08-15 · 15 sites · single snapshot

Sample and method: the same 15 homepage fetches used for our rendering measurement, re-examined for JSON-LD. One request per site with the OAI-SearchBot user agent on 15 August 2026, counting application/ld+json blocks present in the delivered HTML.

Most homepages carry no structured data at all. Of the 14 sites that returned a page, 9 delivered zero JSON-LD blocks. The five that did carried between one and three. These are companies with dedicated web teams, and the majority are shipping homepages with nothing machine-readable about who they are.

How we measured it

We counted occurrences of application/ld+json in the HTML as delivered — not as rendered. That distinction matters: a block injected by JavaScript would not be counted here, and would equally not be seen by any crawler that does not render.

We did not validate the contents. A site with three blocks might have three broken ones. This measures presence only, which is the weakest possible claim and still turns out to be interesting.

Most homepages carry no structured data: the count

Blocks present in the delivered HTML, highest first.

Why this is worth knowing

Structured data is the only part of a page that states, unambiguously and without interpretation, who published it and what it is. Everything else an engine has to infer from prose.

The practical read is not "these companies are doing it wrong". It is that the bar is lower than the volume of advice suggests. If most well-resourced homepages carry nothing, a small site with one accurate Organization block is not behind — it is ahead of most of this list on the one signal that costs an hour.

The two types most of the advice is about are dead anyway

Worth pairing with this: HowTo structured data was deprecated in September 2023 and FAQPage was removed from Google Search in June 2026. Neither appears in Google's current list of supported rich result types. A large share of published schema advice is about two things that no longer produce anything in Google.

That is covered properly, with the four blocks worth actually adding, in schema markup for AI search.

What this does not show

Presence is not correctness, homepages are not the whole site, and one request is not a pattern. Several of these companies almost certainly carry structured data on their product or article templates while leaving the homepage bare — which would be a defensible choice, and this measurement cannot see it.

We also cannot say whether any of it affects citation. No engine publishes how it weighs structured data, and we have not run a controlled test. Anyone who tells you the percentage is guessing.

Common questions

How did you measure this?

One curl per homepage with the OAI-SearchBot user agent, then a count of application/ld+json occurrences in the response body. Two lines of shell. The same fetch produced the readable-text numbers in what AI crawlers get from JavaScript sites.

Does zero blocks mean the site has no structured data anywhere?

No. It means the homepage delivered none. Article and product templates are the more common place to put it, and this sample did not look at those.

Which single block should a small site add?

Organization on the homepage, with sameAs pointing at verified profiles. It is the block that pins down entity identity, which is the fact engines most need and most often have to guess.

The one number to take away

Nine of fourteen well-known homepages delivered zero structured data. The floor here is low, and clearing it costs an hour — writing the block, publishing it and pushing the update for indexing.