Data study · Niche schema types AI answers can read
1,999 Top-1M Sites Still Carry data-vocabulary.org Markup, Six Years After Google Stopped Supporting It
Sites with the retired data-vocabulary.org vocabulary on a crawled page fell from 6,644 in October 2023 to 1,999 in October 2026 (-69.9%). Almost all of it is microdata, and 27.1% of affected pages also carry schema.org microdata.
Analysis of 1,377,729 pages from top-1M sites · HTTP Archive, October 2026, with trend back to 2023
Key findings
- 1
1,999 top-1M sites still had data-vocabulary.org markup on a crawled page in October 2026, on 2,522 pages.
- 2
That is down 69.9% from 6,644 sites in 2023, with a steady drop each year (4,596 in 2024, 3,305 in 2025).
- 3
100.0% of affected pages use it in microdata; JSON-LD and RDFa uses have all but disappeared.
- 4
27.1% of affected pages (684 of 2,522) mix data-vocabulary.org microdata with schema.org microdata on the same page.
- 5
Only 5 sites in the top 10,000 still carry it, against 1,890 sites ranked 100,001 to 1M.
- 6
Forum software stands out: 10.5% of crawled phpBB pages and 5.2% of XenForo pages carry the old vocabulary, against 0.2% of WordPress pages.
Summary
Picture a family-run hardware store that had its website built in 2014. The developer added structured data using the vocabulary that was standard at the time, data-vocabulary.org. The site still works, but that markup has done nothing for search results since 2020, and no AI engine is built around it.
data-vocabulary.org was an early Google-backed vocabulary for structured data. Google stopped supporting it for rich results in 2020 and pointed everyone to schema.org. Pages that still use it carry markup that the main search engine has said it will ignore.
Cleanup is happening, slowly. In October 2026, 1,999 top-1M sites still had data-vocabulary.org markup on a crawled page, down from 6,644 in 2023.
What we measured
We used four October crawls from HTTP Archive (2023 to 2026), mobile client, top-1M sites. HTTP Archive crawls each site's homepage plus one inner page. For each page we checked the rendered microdata item types, the RDFa vocabularies and the JSON-LD script text for any data-vocabulary.org reference.
We counted pages and distinct sites per crawl, recorded whether the page also uses schema.org microdata, and broke 2026 results down by site rank and by technology detected on the page. Platform counts show who still uses the vocabulary, not what any platform outputs by default.
Legacy markup is fading, but slowly
Each year the number of sites falls by roughly a third.
Count
| Crawl | Pages crawled | Pages with data-vocabulary.org | Sites with data-vocabulary.org | Homepages with it | Pages also using schema.org microdata |
|---|---|---|---|---|---|
| Oct 2023 | 1,639,112 | 8,352 | 6,644 | 1,734 | 2,698 |
| Oct 2024 | 1,516,026 | 5,788 | 4,596 | 1,251 | 1,747 |
| Oct 2025 | 1,459,897 | 4,144 | 3,305 | 915 | 1,187 |
| Oct 2026 | 1,377,729 | 2,522 | 1,999 | 572 | 684 |
Source: aisearchoptimization.net analysis of 1,377,729 pages from top-1M sites · HTTP Archive, October 2026, with trend back to 2023
The data cannot separate sites that removed the markup from sites that dropped out of the top 1M.
It lingers on smaller sites
By site rank, 2026.
| Rank bucket | Pages crawled | Pages with it | Sites with it | Page share % |
|---|---|---|---|---|
| Top 10,000 | 13,363 | 6 | 5 | 0.04 |
| 10,001 to 100,000 | 117,602 | 136 | 113 | 0.12 |
| 100,001 to 1M | 1,246,764 | 2,380 | 1,890 | 0.19 |
Source: aisearchoptimization.net analysis of 1,377,729 pages from top-1M sites · HTTP Archive, October 2026, with trend back to 2023
Sites ranked 100,001 to 1M account for 1,890 of the 1,999 sites that still carry it.
Who still uses it
Pages carrying the old vocabulary by technology detected on the page, 2026. A page can run more than one technology.
| Technology | Pages crawled | Pages with it | Sites with it | Page share % |
|---|---|---|---|---|
| WordPress | 210,382 | 389 | 364 | 0.18 |
| XenForo | 3,044 | 157 | 80 | 5.16 |
| phpBB | 1,271 | 133 | 69 | 10.5 |
| 1C-Bitrix | 6,914 | 114 | 101 | 1.65 |
| DataLife Engine | 3,282 | 78 | 66 | 2.38 |
| PrestaShop | 12,242 | 61 | 55 | 0.5 |
| Drupal | 22,234 | 51 | 32 | 0.23 |
| Shopify | 91,943 | 47 | 44 | 0.05 |
| Magento | 14,664 | 29 | 25 | 0.2 |
| WooCommerce | 41,154 | 27 | 27 | 0.07 |
| Blogger | 5,737 | 24 | 13 | 0.42 |
| TYPO3 CMS | 4,971 | 13 | 11 | 0.26 |
| OpenCart | 4,203 | 13 | 12 | 0.31 |
| Joomla | 6,956 | 8 | 6 | 0.12 |
| Wix | 1,529 | 0 | 0 | 0 |
| Squarespace | 878 | 0 | 0 | 0 |
| Webflow | 4,687 | 0 | 0 | 0 |
Source: aisearchoptimization.net analysis of 1,377,729 pages from top-1M sites · HTTP Archive, October 2026, with trend back to 2023
Forum software shows the highest page shares. None of the crawled Wix, Squarespace or Webflow pages carry it.
What this means
View the source of a few pages on your site and search for data-vocabulary.org. If you find it, replace it with the schema.org equivalent, ideally as JSON-LD. Check theme templates and old plugins, since that is often where it hides, and test the result in a schema validator.
Outdated markup does not help search engines or AI assistants understand your pages, and it is a sign that other parts of the site may be due for a technical review. A free AI visibility audit will flag legacy markup along with the gaps that matter more for AI Overviews SEO.
Limits
HTTP Archive crawls one homepage and one inner page per site, so site counts understate the true number of sites with legacy markup somewhere.
The top-1M list changes between crawls, so some of the decline is turnover, not cleanup.
We detect the vocabulary by its URL in item types, RDFa vocabularies and JSON-LD text.
Technology detection depends on HTTP Archive's fingerprints.
Embed this research
Add the charts from this study to your page. The embed links back here.
<iframe src="https://aisearchoptimization.net/embed/data-studies/data-vocabulary-legacy-markup-still-in-use" title="1,999 Top Sites Still Use Markup Google Retired in 2020" width="100%" height="720" style="border:0" loading="lazy"></iframe>
Cite this study
APA
Grenyo, K. (2026). 1,999 Top Sites Still Use Markup Google Retired in 2020. AI Search Optimization. https://aisearchoptimization.net/resources/data-studies/data-vocabulary-legacy-markup-still-in-use
Plain link
1,999 Top Sites Still Use Markup Google Retired in 2020 - AI Search Optimization - https://aisearchoptimization.net/resources/data-studies/data-vocabulary-legacy-markup-still-in-use
Frequently asked questions
Is data-vocabulary.org markup still supported?
No. Google stopped supporting data-vocabulary.org for rich results in 2020 and recommends schema.org instead.
How many websites still use data-vocabulary.org?
1,999 top-1M sites had it on a crawled page in October 2026, down from 6,644 in October 2023.
What should I replace data-vocabulary.org with?
The schema.org equivalent of each type, preferably written as JSON-LD. Remove the old microdata attributes once the new markup validates.
Which sites still use the old vocabulary?
Mostly long-tail sites: 1,890 of the 1,999 sites rank between 100,001 and 1M. By technology, it shows up on 10.5% of crawled phpBB pages and 5.2% of XenForo pages.
Related studies
- Niche schema types AI answers can read1 in 68 JSON-LD Blocks on Top Sites Fails to Parse
- Niche schema types AI answers can readOnly 24.5% of Car Dealer Sites Mark Up Their Vehicles
- Niche schema types AI answers can read28% of Recipe Pages Leave Cook Time Out of Their Markup
- Niche schema types AI answers can read98.8% of Sites Using Hiring Software Lack Job Markup
See where AI search leaves you out
Run the free AI visibility audit. If you want the fixes done for you, the plan is a flat $299/month.