In Hongtao Ren’s hand-selected sample, 32 of 61 reachable homepages returned at least one JSON-LD block in their initial HTML; 29 returned none. That is a description of those homepages’ server responses on one day—not an estimate of how many websites use structured data, and not a test of whether JSON-LD improves rankings.
What the survey counted
Ren says he fetched homepages for 78 well-known domains, selected by hand to span categories, on September 23, 2026. Sixty-one returned a usable page: 42 through a direct connection and 19 through a local proxy. The analysis extracted <script type="application/ld+json"> blocks from each response. It therefore counted markup in the initial server response, not necessarily markup added later by JavaScript in a browser.
As an Amazon Associate I earn from qualifying purchases.
The publisher says the article includes the extraction script and domain list. The counts below are the author’s reported results; they have not been independently re-extracted here. The sample contains reachable homepages only, not inner pages, and was not selected to represent the web as a whole. Ren’s SerpPrism article
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →What appeared in the initial HTML
Of the 61 usable responses, 29 contained no JSON-LD block and 32 contained at least one. “No JSON-LD” here means no block was found in that response. A page could still add structured data during browser rendering, and the result says nothing about that site’s other pages.
#1 Best Overall
Among the 32 responses with JSON-LD, the reported block counts were:
| Blocks in the initial response | Homepages |
|---|---|
| 1 | 24 |
| 2 | 4 |
| 3 | 3 |
| 4 | 1 |
Thirteen of the 32 markup-bearing sites used a top-level @graph. Stripe had the largest reported graph, with 54 nodes. Those are observations about this dataset, not evidence that one graph shape or node count performs better.
Which types and defects did the author report?
Ren classified 20 of the 32 markup-bearing sites as using only generic node types. The following figures are reported type occurrences in the dataset, not a recommended list or a count of sites that ought to use each type:
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →| Type | Reported occurrences |
|---|---|
PostalAddress |
33 |
Organization |
25 |
Place |
24 |
WebSite |
20 |
ImageObject |
20 |
BreadcrumbList |
6 |
SoftwareApplication |
6 |
VideoObject |
6 |
In its 32 markup-bearing responses, the article reports no JSON parse failures, one dangling @id reference, and no site with more than four blocks. These small-sample counts do not establish how common such issues are across websites.
Why 29 of 61 is not a web-wide adoption rate
The domains were hand-selected, the sample was limited to reachable homepages, and the capture was a single-day snapshot. The extraction also looked only at the initial response: JavaScript-injected JSON-LD may not appear there. Ren’s “floor, not a rate” wording refers to that possibility of missing rendered markup; it does not make this non-representative sample a lower-bound estimate for the web.
A broader but differently constructed benchmark offers context, not a direct comparison. The HTTP Archive Web Almanac’s 2024 SEO chapter reports structured data on 48% of desktop homepages and 49% of mobile homepages; for JSON-LD specifically, it reports 41% on desktop and 40% on mobile. These figures come from a broad crawl with its own methodology, conducted two years before Ren’s snapshot. They neither confirm nor contradict the 29-of-61 result as though both measured the same population in the same way. HTTP Archive Web Almanac 2024 SEO chapter
What JSON-LD does—and what it does not prove
JSON-LD is one format for expressing structured data that search engines can interpret. Google lists it as its recommended format among those it supports for rich-result eligibility. Markup must describe relevant content that is visible to readers and satisfy Google’s policies; correct markup still does not guarantee a rich result. Google’s structured data introduction and General structured data guidelines
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
That eligibility is distinct from ordinary web ranking. Google says structured-data issues that trigger a manual action affect eligibility for rich results, not ordinary search ranking. Ren’s homepage survey measured markup presence and patterns; it did not compare rankings or test whether adding JSON-LD changes them.
Should I remove SearchAction markup?
Not because of an assumed ranking penalty. Fourteen of Ren’s 32 JSON-LD responses contained SearchAction markup associated with Google’s former sitelinks search box. Google announced on October 21, 2024, that it would remove that search feature globally starting November 21, 2024. Google said the change did not affect rankings, site owners did not need to remove the markup, and unsupported markup of this kind would not cause Search or Search Console errors. John Mueller wrote, “While you can remove sitelinks search box structured data from your site, there’s no need to do so.” Google Search Central: Farewell, Sitelinks Search Box
Rank #4
- How search engines work: Because knowing your enemy is half the battle.
- SEO Basics: Like how to start a website and submit it to Google.
- Keyword Research: Find the most promising keywords for your business.
- SEO Content: Create content that even search engines want to binge-read.
- On-Page SEO: The most effective way to explain your pages to search engines.
Remove or retain the legacy block according to your maintenance preference: removing it can make a schema graph less confusing, while retaining harmless legacy markup avoids unnecessary work. This advice concerns the retired sitelinks search box markup, not all WebSite structured data.
Is @graph required?
No. Ren found top-level @graph in 13 of the 32 markup-bearing responses, but the survey did not compare outcomes for sites with and without it. It is a way to organize multiple linked nodes in JSON-LD, not a demonstrated ranking advantage or universal requirement. Use a graph when it makes the relationships in your data clearer; otherwise, a valid JSON-LD structure that accurately describes the page matters more than matching the sample’s pattern.
Do not confuse the graph pattern with whether a particular Google feature supports the types and properties you provide. Check the current documentation for the feature you want rather than copying types from another homepage.
Best Value
Do I need structured data to rank?
This survey cannot establish that. Its counts describe markup in homepage responses, not ranking performance, and Google’s documentation frames structured data around eligibility for supported search features rather than a promise of better ordinary rankings. Add it when it accurately represents useful page content and supports a current feature or clear data model—not simply to imitate sites in this sample.
A practical homepage markup audit
- Check what the page actually says. Compare each node and property with the homepage’s visible content and the organization or site information it represents. Remove inaccurate or irrelevant claims.
- Check the intended Google feature. Consult the current documentation for the relevant structured-data type and confirm that the page and markup meet its policies. A validator can catch syntax problems, but passing a test does not promise display.
- Inspect both response and rendered page when needed. If markup is generated by JavaScript, a raw-response extraction can miss it. Compare the initial HTML with the browser-rendered DOM before concluding that the page has no JSON-LD.
- Resolve references. For each
@idthat points to another node, confirm that the referenced identifier is defined where expected and that the relationship is intentional. - Decide whether to keep legacy markup. For retired sitelinks search box
SearchAction, base removal on clarity and maintenance—not fear of a ranking impact Google says does not occur.
For site names, do not assume that all WebSite markup is obsolete: Google’s current documentation describes homepage WebSite structured data as one way to indicate a preferred site name. Google determines the displayed name automatically and may use other sources as well. Google documentation on site names
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
