Google Killed FAQ Rich Results on May 7. The March Core Update Told You What to Build Instead.

person

AI SEO Intelligence

calendar_today May 21, 2026

schedule 11 min read

Google Killed FAQ Rich Results on May 7. The March Core Update Told You What to Build Instead.

On May 7, 2026 Google officially stopped rendering FAQ rich results in search. Weeks earlier, the March 2026 Core Update finished rolling out - and across the SEO post-update analyses we read, one signal kept coming up as the most consistent winner: firsthand Experience. The two events landed close together, and they reward and penalize content in the same direction. The first removed a cheap structured-data optimization that used to earn a SERP feature. The second moved weight onto the kind of content that takes a real person doing real work to produce. Same underlying preference, two surfaces.

If your SEO strategy still leans on FAQ schema as a citation lever, or if your E-E-A-T audit still treats "Experience" as a soft tiebreaker rather than a primary input the post-update traffic distribution visibly reflects, this post is for you. We rebuilt our own audit weights last week in response to both shifts. Here is what the data says, and what changed.

Key Takeaways

Google stopped rendering FAQ rich results on May 7, 2026. Search Console drops the FAQ report in June. API support ends in August. The markup remains valid Schema.org but no longer earns a SERP visual.
FAQ schema is now dead on both surfaces - no rich result (Google) and no measurable AI citation lift (May 2026 study of schema's impact on AI citations, via Search Engine Roundtable).
The March 2026 Core Update boosted authoritative, brand-owned, and first-party content; surface-level summaries and aggregator content lost ground.
Original research and firsthand Experience are now functionally co-equal pillars of "information gain" - we rebalanced our audit to reflect this.
The optimization effort you save by stopping FAQ schema maintenance is exactly the effort you need to redirect into Experience signals.

The FAQ Rich Result Timeline

Google's deprecation rolled out faster than most retirements. The official sequence:

May 7, 2026 - FAQ rich results stopped rendering in Google Search. SERPs no longer show expandable Q&A blocks under FAQ-marked pages, even for grandfathered government and health domains that had retained eligibility after the 2023 narrowing.
June 2026 (announced) - The FAQ enhancement report is removed from Google Search Console.
August 2026 (announced) - Search Console API support for FAQ structured data is dropped entirely.

The markup itself remains valid Schema.org. Google will not penalize you for keeping FAQPage JSON-LD on your pages. But the schema no longer produces a search feature, no longer surfaces in Search Console, and - as of August - no longer registers as a tracked enhancement. It is metadata for an outcome that no longer exists.

Worth pausing on: this is the third rich result deprecation in three years. HowTo was restricted to desktop in 2023 and dropped entirely from mobile. FAQ was narrowed to "well-known authoritative" domains in 2023. Both were then fully retired in 2026. The pattern is consistent with a Google preference for retiring features that were widely spammed and whose click-through impact stayed below whatever bar the company maintains internally - though Google itself has never published those thresholds, so the inference is reading tea leaves with the leaves removed.

Schema Is Dead in Both Directions

If FAQ schema were dead only on the Google SERP, you could argue it still earned its place as an AI search signal. Many of the "AI SEO" guides that proliferated through 2025 made exactly that pivot: yes, FAQ rich results are weakening, but ChatGPT and Perplexity tokenize the JSON-LD and cite it preferentially. Add FAQPage schema and your AI citation rate climbs.

The data published ten days ago killed that argument. We covered the AI-citation evidence in detail in our companion FAQ post; the short version: a May 2026 study reported by Search Engine Roundtable found adding schema markup did not significantly increase pages' likelihood of being cited in AI Overviews or ChatGPT. SEJ's parallel analysis reached the same conclusion: schema does not directly influence AI Overview citation selection - quality and authority do. The widely-circulated "FAQ schema lifts AI citations by 30 percent" claim does not survive scrutiny.

Put the two findings next to each other and FAQ schema becomes a maintenance burden with no measurable SEO or AI-citation upside. Google will not give you a rich result. AI engines will not give you a citation lift. The schema still validates, the JSON still parses, the page still passes Schema.org checks - there are also legitimate non-ranking reasons to keep the markup (entity disambiguation, internal content governance, possible future reuse), and we're not arguing for ripping it out. We're arguing it should no longer be a workflow priority you actively expand or new content actively adds.

The optimization is technically alive but the ranking and citation outcomes it was supposed to produce no longer exist. Teams will spend another year maintaining FAQPage blocks under the assumption the schema is doing SEO work, when the SEO work moved elsewhere months ago.

What Quietly Changed in the April Core Update

While the FAQ rich result news drew most of the SEO press, the March 2026 Core Update finished rolling out and produced a much more useful signal in post-update analyses.

Multiple post-update analyses converged on the same structural finding. Search Engine Land's coverage found the March 2026 update "more volatile than December" with authority signals gaining weight - official sites, brand-backed content, and data-rich sources boosted while aggregators and low-quality content sites dropped. Amsive's analysis by Lily Ray described it as a "first-party, official-source correction": Google tilted toward authoritative, brand-owned, and government domains while penalizing UGC platforms and SEO-visibility-first content. The pattern pointed to the Experience signal - the extra "E" Google added to E-E-A-T in December 2022 - as a load-bearing factor: pages demonstrating genuine first-hand work gaining ground, surface-level summaries and aggregator content losing it.

We don't have a direct view into Google's algorithm. What we have is consistent third-party measurement of which pages gained and which lost after the March rollout - and the Experience signal moved from a thing quality raters scored to a thing the post-update traffic distribution visibly reflected. Read it as a strong inference from observed rankings, not as confirmed internal mechanics. The signals that mark "the author has done this thing" are now load-bearing in those analyses:

First-person specifics: "I tried," "after 6 months," "the difference was"
Process narratives: walking through what you did, what failed, what you changed
Photos that are not stock photos: original screenshots, real workspace, actual artifacts
Numbers from your own measurement, not aggregated from other sources
Acknowledged limits: what the experience does not generalize to, where it might not apply

What it is not: an author bio with credentials, a publish date, or a byline link to a profile page. Those signal Expertise and Authority - two of the older three letters. Experience is the harder one because it requires the author to actually have done the thing, and the article to mention it concretely enough to be verifiable from the text itself.

Why Both Shifts Reward the Same Kind of Content

The pattern across both events is the same: Google and AI engines are getting better at telling the difference between content that describes a topic and content that demonstrates knowledge of it.

FAQ schema was the canonical example of the first category. It is metadata about content. It declares "here are questions and answers" without producing any new information the user could not get from reading the page. Google's rendering pipeline used to convert that metadata into a rich snippet; once the snippet stopped rendering, the metadata's only ranking job evaporated.

The March 2026 Core Update rewards the second category. It rewards content where the author's hands are on the topic. Both shifts point in the same direction - removing weight from a low-information signal, adding weight to a high-information one. The February 2026 Discover Core Update pushed the same way on the Discover surface a few months earlier: aggregator publishers lost ground, original-reporting publishers gained, and the algorithmic reward for thin, structurally-optimized content dropped at the same time as it rose for genuine reporting.

The trade-off is not effort-for-effort. Adding FAQPage JSON-LD is cheap; sourcing original photos, running a real test, writing about a failure mode you actually encountered is not. The honest implication is that the easy SEO work just lost its outcome, and the work that still produces outcomes is the work that takes a real subject-matter operator to do. If your editorial budget was leaning on the cheap side, that's the part that has to grow.

What We Changed in How We Score Pages

We recalibrated our scoring last week in response to both events.

FAQ schema is no longer a positive signal in our score. We used to give pages a small structured-data bonus when FAQPage JSON-LD was present, and we used to warn FAQ-intent pages that lacked it. Neither of those movements survives the May 7 deprecation or the schema citation study evidence. FAQ markup is now observed but not weighted: the audit still surfaces that the schema is on the page (so teams migrating off it can verify), but the score no longer moves whether you have it or not. Article-type pages still get the same warning if they lack the Schema.org types Google still grants rich results to.

The cost of missing firsthand experience went up. Our content-quality scoring asks, in effect, whether an article demonstrates that the author has personally done what it's writing about. When the answer is clearly no, the page's score takes a hit on the AI-search dimensions. After the March 2026 Core Update, that hit looked too small relative to what the post-update traffic distribution implied. We raised it to match the weight already assigned to missing original research. The two signals now carry equal weight in the score, because the Q2 2026 evidence treats them as equally load-bearing.

That second change matters more than it sounds. Most "content quality" scoring systems still treat Experience as a tiebreaker - a small adjustment after the bigger signals are computed. The March rollout suggests that's understating it. Pages that surface concrete firsthand specifics (numbers from your own measurement, photos of your real workspace, walkthroughs of failure modes you actually hit) are visibly winning against pages that don't. Audit tools that price the signal as a tiebreaker are now mis-scoring the gap.

What we did not change: the big penalty on missing structured data altogether. The schema citation study evidence argues for lowering it on AI-citation grounds, but the current evidence base isn't enough to move our largest non-indexing penalty. We're holding for Q3 replication evidence before touching that one.

A 10-Minute Self-Audit for Your Top Pages

Pull up your top five revenue-driving article-type pages. Walk each one through this:

Does the page have FAQ schema? If yes - leave it. Do not invest more in it. Stop expanding the FAQ block in the schema beyond what the visible page already has. The maintenance work is no longer producing ranking outcomes.
Does the visible page have a real FAQ section? Three or more question-format H2 or H3 headings, with self-contained answers. (Our sister post walks through the FAQ patterns AI engines actually cite - the visible structure does the work the schema doesn't.)
Read the first 300 words. Can you identify the author's firsthand experience with the topic? Look for first-person specifics, process narratives, original numbers. This is an audit heuristic, not an algorithm probe - but it correlates well with the post-update analyses, and it's the same signal our scoring now penalizes more heavily when it's missing.
Check the images. Are any of them original screenshots, photos from the actual context, or annotated diagrams the author created? Or are they all stock images and AI-generated illustrations?
Check the numbers. Does the article cite original measurements (you ran X across Y samples) or aggregated industry stats (according to Statista, 70% of marketers...)? Both have a place. The Q2 2026 core update evidence rewards the first.
What does the article admit it doesn't know? Pages that acknowledge limits, edge cases, and "this didn't work for me in scenario X" score higher on Experience signals than pages that present uniform confidence. Counter-intuitive, well-documented.

If your top five fail steps 3 through 6 across the board, you have the same problem half the web has after the March Core Update: optimized for the older E-E-A-T (Expertise, Authoritativeness, Trust) while the algorithm has moved its weight onto the newer one (Experience).

What This Means for Your Roadmap

The deprecation + the core update together suggest three editorial-process changes worth queuing for Q3 2026:

Stop expanding FAQ schema. Audit your FAQPage blocks once for accuracy, then freeze maintenance. The hours you spend keeping schema in sync with FAQ section edits are now producing no measurable ranking or citation lift. Redirect those hours.

Add an "Experience" section to your editorial workflow. Before publication, the writer is asked to flag the specific firsthand elements the article contains. If the answer is "none," the article doesn't go out - or it goes out tagged as an aggregator piece that nobody on the team should be surprised when it doesn't rank.

Re-audit your existing top-traffic articles. Use the 10-minute checklist above. The articles that won pre-March 2026 updates were not necessarily Experience-strong; they were SEO-fundamentals-strong. Some of them are about to lose visibility on the next core update if they aren't refreshed. Better to find the gap now than after the next traffic drop.

The two Q2 2026 shifts are not catastrophic individually. Together they tell a coherent story about where Google and AI search are headed: metadata is being deprecated, demonstration is being rewarded. The audit weights, the editorial workflow, and the optimization budget should all move in the same direction.

Want to see how your pages score against the Q2 2026 weights? Run a free audit at hybridranking.com - firsthand-experience and structured-data signals are part of the 99 we evaluate, freshly calibrated to the March Core Update and the May 7 deprecation. If you use Claude Desktop or Cursor and want the current weights surfaced in-conversation rather than via training-data snapshots, our MCP server ships exactly that integration.