Googlebot’s JSON-LD Extraction Change: What Website Owners Need to Check

Confirmed: Google has changed how Googlebot extracts JSON-LD from HTML. It now applies one pass of HTML unescaping and follows standard JSON formatting more strictly. As a result, structured data that has been HTML-escaped more than once may not be interpreted correctly.

JSON-LD is often embedded in a page inside a <script type="application/ld+json"> element. If a publishing system, template, plugin or rendering process unnecessarily escapes characters in that block—and another layer escapes them again—the final markup delivered to Googlebot can be double-escaped.

Our read: This is less a change to schema vocabulary than a stricter handling of how JSON-LD is encoded. Correctly formatted JSON-LD should continue to work. The immediate risk is concentrated on sites whose content pipelines transform schema markup before it reaches the final HTML.

Structured data helps search engines understand entities, page types and relationships. It can also make eligible pages available for supported search features. If Googlebot cannot parse a JSON-LD block, the markup may provide no usable signal even when its underlying schema properties are otherwise correct.

A page can look normal to visitors while serving broken structured data in its source or rendered HTML. That makes this issue easy to miss during visual quality assurance. It may also affect many pages at once when the error comes from a shared template, tag manager, plugin or server-side rendering process.

Confirmed: Double-escaped structured data may need to be corrected so Google can read it properly. It is not yet known how many sites are affected, and the update does not mean that valid structured data guarantees a rich result or other search feature.

For SEO, malformed JSON-LD can reduce Google’s ability to use structured information associated with products, articles, organizations, events and other supported types. This does not necessarily remove a page from the index, but it can affect eligibility for relevant enhancements and weaken machine-readable context.

For AEO and GEO, clean structured data can reinforce explicit facts about an entity or page. It is not a shortcut to inclusion in an answer engine or generative response, and there is no confirmed direct link between this extraction change and AI citations. However, markup that cannot be parsed cannot contribute any machine-readable context to systems that rely on Google’s understanding of the page.

For SXO, the impact is indirect but meaningful. Search enhancements can help users understand a result before clicking, while accurate page classification can support a clearer search journey. Structured data must still match visible content; fixing encoding does not compensate for incomplete, misleading or low-quality information.

For content strategy and discoverability, teams should treat schema as part of the publishing workflow rather than a one-time technical addition. Changes to a CMS, JavaScript framework, consent platform or deployment pipeline can alter JSON-LD without changing the visible page. Important templates and high-value content therefore need recurring validation.

Start with pages that use structured data at scale, especially templates connected to revenue, lead generation, publishing or local visibility. Test both the HTML initially returned by the server and the rendered DOM because the problem may be introduced at different stages.

Do not simply replace every HTML entity you find. JSON strings can legitimately contain escaped characters. The goal is valid JSON-LD after the single HTML-unescaping pass described in the update, not the removal of all escaping.

CrawlWeb can support the audit process by helping teams review website templates, technical SEO issues and page groups at scale rather than checking URLs one by one. Audit findings can be prioritized alongside broader SEO, AEO and GEO considerations so teams can see where broken machine-readable markup overlaps with important content.

Search Console insights can help identify affected page types and track relevant search changes after a correction. Competitor research and content-opportunity analysis can also provide context for discoverability, but they cannot prove that a competitor receives AI citations because of structured data.

CrawlWeb should be used alongside direct validation of the final JSON-LD delivered to Googlebot. A crawler or audit report can identify patterns and prioritize investigation, while Google’s testing tools and Search Console remain important for confirming how Google processes individual pages.

Search Engine Roundtable reported the extraction change and its implications for double-escaped structured data: https://www.seroundtable.com/json-ld-extraction-googlebot-41921.html