Structured data used to be the quiet part of technical SEO: add a chunk of JSON-LD, hope for a rich snippet, move on. In 2026, the rules are different. Google is tightening how it parses markup, trimming which rich results still exist, and pushing the real value toward product and policy data that machines can actually use.
Three changes you cannot ignore
The first is crawl weight. In February 2026, Google cut Googlebot’s fetch ceiling from 15MB to 2MB per URL. At that threshold, Google truncates rather than rejects, so JSON-LD placed deep inside a very large HTML file may not be seen at all.
The second is parsing. On 21 August 2026, Googlebot moved to a single pass of HTML unescaping for JSON-LD. Markup that double-encodes entities can break even when the vocabulary is otherwise perfect.
The third is feature removal. FAQ rich results stopped appearing on 7 May 2026. Seven other structured data types were deprecated in June 2025, with Google product manager Henry Hsu saying the affected displays “no longer added significant value.”
Commerce is where schema pays off
While informational formats contract, commercial markup keeps expanding. Product variant support arrived in February 2024 through hasVariant and variesBy. Return policies gained a mandatory two-letter country field in March 2025. Loyalty programmes can be expressed with MemberProgram, and shipping policy markup opened to merchants without Merchant Center accounts in November 2025. In July 2026, Google extended Product category values to accept CategoryCode objects, making on-page markup align more cleanly with feed data.
This pattern makes sense. Search engines no longer need structured data to parse an FAQ or a definition. They do need explicit price, currency, availability and policy fields for shopping agents, AI assistants and merchant experiences.
The ranking myth is still a myth
There is a persistent belief that schema boosts rankings. It does not. Valid markup makes a page eligible for a rich result or extraction, not more authoritative. Ahrefs tested schema across pages already appearing in AI Overviews and found no measurable ranking difference. Consultant Marie Haynes cited that result when arguing that informational schema deserves little attention, while commerce data deserves much more.
There is also enforcement. A structured data manual action can leave a page indexed while Google ignores its markup, removing the display benefit without a traffic collapse to signal the problem. Undisclosed incentivised reviews moved into that enforcement category on 24 July 2026.
How to use this
For most marketing and growth teams, the task is not more schema types. The first public usage dataset, released in June 2026, shows how concentrated adoption is: twelve core types sit above ten million domains, while around half the vocabulary is effectively unused.
A practical framework is Check, Align, Reduce.
- Check that key product, article and organization pages carry valid JSON-LD.
- Align on-page schema with Merchant Center data so price and availability do not conflict.
- Reduce page weight so Googlebot’s 2MB fetch ceiling does not hide your markup.
Publishers should not abandon schema either. Google confirmed in March 2026 that schema.org and og:image both feed thumbnail selection in Search and Discover. Apple uses the schema.org isAccessibleForFree property as a paywall signal, keeping pages indexable while excluding them from AI-generated output. ProfilePage and DiscussionForumPosting remain valid options for community and editorial teams.
Structured data is not a growth hack. It is infrastructure. Get the commercial basics right, keep the pages light, and let the rich results follow.
Source: PPC Land
