Schema’s SEO Value May Be Dwindling — But It’s Got One Thing Going For It
“Structured data isn’t dead. It’s just not doing what you think it’s doing anymore.” — Iyer, Founder of ACME.BOT
Schema markup was once known as a lever, where adding structured data resulted in rich results that drove traffic on your site. But that trade-off has collapsed. Google’s now keeping your JSON-LD data to feed its LLMs while simultaneously removing the visible reward from the table. From May 2026, you won’t see FAQ rich results.
Here’s When FAQ Rich Results Disappeared
Google didn’t flip a switch one day and decide to remove FAQ Rich Results. They first restricted some, then moved to remove them completely.
On May 7, 2026, FAQ rich results stopped appearing in Google Search across the board. No exceptions. Not even for government and health sites that had been previously exempt. But here’s the thing: the March 2026 core update had already cut FAQ impressions by roughly 50%, so this wasn’t a surprise—it was a confirmation.
The rest proceeded in typical fashion:
`
| Date | Action |
|---|---|
| May 7, 2026 | Stop showing FAQ rich results in searches |
| June 2026 | Remove support for FAQ search appearance, rich result report and Rich Results Test |
| August 2026 | Stop supporting FAQ rich result in Search Console API |
Google even tells us in their own words that they’ll “continue to use FAQ structured data to better understand pages” This means the feature might go away, but your markup will still live on. You’re not losing the schema. You’re losing the real estate.
That distinction is the whole game now.
Google’s Been Shrinking Structured Data Rich Results for Years
If this seemed like a sudden change, it wasn’t. Google has been following the same approach for three years — restricting and then removing structured data rich results. Here’s the receipts.
| Date | Action | Impact |
|---|---|---|
| August 2023 | Locked “How-To” to desktop & reduced FAQ visibility. | First signal |
| September 2023 | Removed How-To rich results from all devices. | Restriction → full kill |
| January 2025 | Removed breadcrumbs from mobile. | Another quiet casualty |
| May 2026 | Removed FAQ rich results from all devices. | Pattern complete |
No surprise to you that Google is wrapping up the feature. We’ve seen them tighten the visibility screws, wait a few weeks to a few months, then shut it down completely. How-To went from “desktop only” to “deprecated” in about three weeks. Breadcrumbs then followed a similar path. And now we have FAQ.
This isn’t simply a one off thing. It’s the culmination of a multiyear pullback. The question isn’t whether this pattern stops. It’s why Google is even doing this in the first place. And that impacts how you should approach your schema right now.
Why This Is Happening: LLMs Don’t Need JSON-LD to Read Your Page
LLMs can look at raw HTML and understand things like entity relationships, content heirarchy, and Q&A organization — even if it isn’t marked up on the page.
Think your FAQ schema is doing the heavy lifting? It’s not. Feed an LLM raw HTML with a question and answer in plain prose — it maps the relationship without touching a single schema tag. Google’s AI Overviews and AI Mode run on these same models. Explicit markup? Increasingly redundant for comprehension.
Here’s one gap that Google still won’t close: understanding a page vs surfacing a rich result. They still say they use FAQ structured data to “understand pages” —which is technically true. However, that hasn’t led to the same cool dropdowns we saw before. We’ve seen ACME.BOT clients spend hours maintaining FAQ schema, while neglecting the actual quality of their content—the actual thing that AI search rewards.
What’s really valuable for an AI search?
- Answer-ready prose that LLMs can cite directly
- Content hierarchy and entity relationships visible in plain text
- Unique, substantive answers that earn LLM visibility over boilerplate
Schema was always a workaround. Now it’s not even that.
What SEOs Should Actually Do Now
Schema markup really isn’t an SEO lever anymore. Here’s what is:
- Stop treating structured data as a growth engine. You’re spinning JSON-LD cycles that don’t surface in Google’s interface. While Google still parses structured data, the rich result is now gone. Move on.
- Keep schema only in places where Google uses it.
| Schema Type | Keep? |
|---|---|
| Product | Yes |
| Event | Yes |
| Recipe | Yes |
| FAQ | No |
- Write prose that is the structure. Q&A styled naturally in a paragraph beats out hiding markup. Models can extract markup relationships derived from human-readable HTML. When it comes to AEO, writing content that answers questions before they’re asked gets LLM visibility and AI citations.
- Focus on unique, answer-ready content. An AI citation isn’t something your schema will solve—it’s more of a content challenge to be addressed. If you want to rank on Google and have your content cited by AI, you need to build a content engine that does just that. The structure AI search is looking for isn’t found within your tags—it’s found in your prose.