SERP Features

TL;DR — SERP featuresSERP features are any element on a search results page beyond the classic ten blue links — featured snippets, People Also Ask, knowledge panels, sitelinks, image and video packs, AI Overviews, and structured-data-driven rich results. Google documents that losing rich-result eligibility doesn't affect ranking; their SEO relevance is CTR (they redistribute clicks). Some are unlocked by markup; most are purely algorithmic. are everything on a Google results page that isn’t a plain blue link — featured snippets, the “People also ask” box, star ratings, image and video rows, knowledge panelsThe Knowledge Graph is Google's database of entities — people, places, organizations, and things — and the factual relationships between them. It's separate from any single website's structured data: your schema markup is one of many possible inputs to the graph, not the graph itself., AI OverviewsAI Overviews are the AI-generated summary box Google shows above or within its regular search results, written by Gemini models from pages retrieved out of Google's normal Search index. It's a Search feature, not a separate platform or index.. Some are unlocked by adding structured dataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. to your page; most are chosen automatically by Google and you can’t switch them on. None of them make you rank higher — they just change how your result looks and who clicks what.

What SERP features are

“SERP feature” is an industry umbrella term for search-result presentations beyond a plain blue link; Google documents individual search galleries rather than one universal SERP-feature taxonomy. Evidence for this claim Google's Search Gallery documents supported search appearances by feature and content type. Scope: Google Search result features; the industry term SERP feature is broader and not one formal Google taxonomy. Confidence: high · Verified: 2026-07-14 Google: Search Gallery Structured data can make a page eligible for supported rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show. but does not guarantee display. Evidence for this claim Valid structured data can establish eligibility for supported rich results but does not guarantee that Google will show them. Scope: Google rich-result eligibility and quality guidance. Confidence: high · Verified: 2026-07-14 Google: Understand structured data

“SERP” stands for search engine results page. For years that page was basically ten blue links. It isn’t anymore. Today almost any search shows extra stuff — a boxed answer at the top, a row of images, “People also ask” questions you can expand, a panel about a company or celebrity on the right, an AI-written summary. All of those extras are SERP features.

One naming note worth knowing: “SERP feature” is what the SEO industry and tools (Ahrefs, Semrush) call them. Google’s own help docs don’t use that phrase — they say “search result features” or “visual elements.” Same thing, just the official name.

The one distinction that matters

SERP features come in two flavors, and mixing them up is the single most common mistake I see:

• Features you can unlock with code. If you add structured data (also called schema markupSchema markup is code that uses the schema.org vocabulary to label what your content means so search engines can understand it and show rich results. It's most often written in JSON-LD, and it's not a direct ranking factor.) to a page, it becomes eligible for a fancier listing — star ratings under a review, a price on a product, breadcrumbsBreadcrumbs are a secondary navigation trail (Home > Category > Page) that shows where a page sits in a site's hierarchy. They create internal links that pass PageRank, and when marked up with BreadcrumbList structured data they can drive the path Google shows in desktop search results.. These are called rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show..
• Features you can’t switch on at all. Things like the featured snippet (the boxed answer), “People also ask,” knowledge panels, and AI OverviewsAI Overviews are the AI-generated summary box Google shows above or within its regular search results, written by Gemini models from pages retrieved out of Google's normal Search index. It's a Search feature, not a separate platform or index. are chosen by Google automatically. There’s no tag you add to get one. Google literally says you can’t request a featured snippet.

Most articles online blur these together and imply markup helps everywhere. It doesn’t. Knowing which bucket a feature is in tells you whether you can do anything about it.

Do SERP features help you rank?

No, not directly. This is the big myth. Google’s own documentation is explicit about the structured-data case: losing rich-result eligibility “doesn’t affect how the page ranks.” That’s a directly documented sentence — for structured data specifically. Google hasn’t published an identical blanket line for every other SERP feature, but the same logic is understood to hold: these are a display layer Google chooses whether to show you, not a separate signal you feed into ranking. SERP features change how your result is displayed, not (as far as Google has actually said) where it sits.

What they do change is clicks. A star-rating listing can pull more clicks than a plain one. But a big AI Overview or a boxed answer at the top can also pull clicks away from everything below it — including you. So the honest way to think about SERP features is: they redistribute attention on the page, sometimes in your favor, often not.

Want the full taxonomy, the structured-data-vs-algorithmic split in detail, the CTR data, and the zero-click debate? Switch to the Advanced tab.

TL;DR — SERP featuresSERP features are any element on a search results page beyond the classic ten blue links — featured snippets, People Also Ask, knowledge panels, sitelinks, image and video packs, AI Overviews, and structured-data-driven rich results. Google documents that losing rich-result eligibility doesn't affect ranking; their SEO relevance is CTR (they redistribute clicks). Some are unlocked by markup; most are purely algorithmic. split cleanly into two mechanisms, and that split is the spine of this whole topic: structured-data-driven rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show. (Product, Recipe, Event, Review — you add markup to become eligible, never guaranteed) vs. purely algorithmic features (featured snippets, People Also Ask, knowledge panels, AI OverviewsAI Overviews are the AI-generated summary box Google shows above or within its regular search results, written by Gemini models from pages retrieved out of Google's normal Search index. It's a Search feature, not a separate platform or index., image/video packs, sitelinksSitelinks are extra links from the same domain that Google clusters together under a single search result, usually for branded or navigational queries. They're generated entirely algorithmically — there's no way to add, edit, or guarantee them. — no markup lever exists). Google is explicit you can’t request a featured snippet. None of it is a ranking factor — Google’s docs say structured-data eligibility “doesn’t affect how the page ranks”. The real reason to care is CTR: features redistribute clicks, and more features on a SERP generally lowers position-1 CTR. The structured-data side keeps shrinking (FAQ, HowTo, six types in June 2025); the algorithmic side keeps growing (AI OverviewsAI Overviews are the AI-generated summary box Google shows above or within its regular search results, written by Gemini models from pages retrieved out of Google's normal Search index. It's a Search feature, not a separate platform or index.).

”SERP feature” isn’t Google’s term

Google’s Search Gallery organizes supported result experiences by feature and content type rather than defining a single formal category called SERP features. Evidence for this claim Google's Search Gallery documents supported search appearances by feature and content type. Scope: Google Search result features; the industry term SERP feature is broader and not one formal Google taxonomy. Confidence: high · Verified: 2026-07-14 Google: Search Gallery Eligibility requirements and actual appearance are separate: valid markup does not guarantee a rich result. Evidence for this claim Valid structured data can establish eligibility for supported rich results but does not guarantee that Google will show them. Scope: Google rich-result eligibility and quality guidance. Confidence: high · Verified: 2026-07-14 Google: Understand structured data

Start with the vocabulary, because it’s a good authority tell. “SERP feature” is industry/tool language. Google’s own docs use “search result features” and “visual elements,” and its Visual Elements Gallery defines them as “the building blocks of the Google Search results page that a user can perceive or interact with.” That gallery is also a cleaner mental model than the flat “13 SERP features” lists you’ll find elsewhere, because it groups everything into five buckets:

• Text result — “a result in Google Search that’s based on the textual content of the page.” The plain blue link.
• Rich result — “a result that typically relies on structured data in the markup of your page to display graphical elements or interactive experiences.”
• Image result — “a result that’s based on an image that’s embedded on that web page.”
• Video result — “a result that’s based on a video that’s embedded on that web page.”
• Exploration features — “a feature that helps searchers expand and refine their initial search” (related searches, People Also Ask–style refinement).

Notice that only one of those five buckets — rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show. — is described as relying on your markup. That’s the whole distinction in Google’s own words.

The spine: structured-data-driven vs. purely algorithmic

Here’s the split every SEO should internalize. On one side are features you can make yourself eligible for with markup. On the other are features Google and Bing decide to show based on the query and the content, with no markup switch at all.

Structured-data-driven (need markup — see the Search Gallery): Product, Recipe, Event, Review snippetReview schema (schema.org/Review) is structured data for a single critic's or user's evaluation of one specific thing — one author, one itemReviewed, one reviewRating — distinct from AggregateRating, which summarizes many reviews into an average., Breadcrumb, and the ~25 other current types. Google uses structured data to understand the content on the page and show it “in a richer appearance in search results, which is called a rich result.” This is exactly the topic of the Rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show. deep dive, and the whole Structured DataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. hub this article lives under. Two hard rules from that hub carry over: markup makes you eligible, never guaranteed — Google “does not guarantee that your structured data will show up in search results, even if your page is marked up correctly” — and the supported list keeps shrinking (more on that below).

Purely algorithmic (no markup lever): Featured snippets, People Also Ask, knowledge panelsThe Knowledge Graph is Google's database of entities — people, places, organizations, and things — and the factual relationships between them. It's separate from any single website's structured data: your schema markup is one of many possible inputs to the graph, not the graph itself., AI Overviews, image and video packs (derived from on-page media, not from schema), local packs, sitelinksSitelinks are extra links from the same domain that Google clusters together under a single search result, usually for branded or navigational queries. They're generated entirely algorithmically — there's no way to add, edit, or guarantee them., top stories, discussions/forums. You cannot tag your way into any of these. Conflating this bucket with rich results is the single most common error in competing articles — they list every feature with “add markup” implementation tips, as if schema were a universal lever. It isn’t.

Are SERP features a ranking factor? Documented for structured data, inferred for the rest

This is worth being precise about, because the parent hub already establishes that structured dataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. isn’t a ranking factor via John Mueller’s repeated statements. There’s an even stronger, documented version for this article — but it’s narrower than “no SERP feature is ever a ranking factor,” and it’s worth being exact about the scope. Google’s structured-data policies page, describing what happens when a page gets a structured-data manual action, says a page “loses eligibility for appearance as a rich result; it doesn’t affect how the page ranks in Google web search.” That’s an actual documented sentence separating rich-result eligibility from web ranking — cleaner than any relayed office-hours paraphrase — but it’s specifically about the structured-data manual-action case. Google hasn’t published an equivalent line for featured snippets, PAA, knowledge panels, or AI Overviews. The reasonable extension — supported by how those systems actually work (a display-layer selection made on top of the existing ranking, not a separate scoring input) — is that the same eligibility-vs-ranking split holds across the board. Treat that extension as informed inference, not an identically documented fact, and lean on the sd-policies sentence as the one case Google has actually put in writing.

The corollary John Mueller has offered is that doing something technically correct on a page doesn’t by itself make it a better page than it would be otherwise — technical correctness isn’t the same as content quality. That Mueller framing is sourced only via Search Engine Journal’s coverage; the companion primary coverage wasn’t fetchable for direct verification, so treat it as secondary color, not a verbatim quote — the documented sd-policies line above is the citation to lean on.

How each major feature actually gets triggered

Featured snippets are the proof point for “algorithmic, not markup.” A featured snippet is a “special box where the format of a regular search result is reversed, showing the descriptive snippet first.” Can you request one, or tag a page as snippet-worthy? Google’s answer is blunt: “You can’t. Google systems determine whether a page would make a good featured snippet for a user’s search request, and if so, elevates it.” The only lever site owners actually have is the opposite one — opting out with nosnippet or a lower max-snippet value. It’s a removal lever, not an acquisition one.

People Also Ask is a case worth flagging for a different reason: there is no dedicated Google Search Central documentation page for PAA, despite it appearing on the majority of English-language SERPs. Every mechanics claim you read about PAA is industry inference from observation, not documented fact. That’s a small but genuine credibility beat — most articles describe PAA authoritatively while citing nothing, because there’s nothing official to cite.

Knowledge panels sit on the spectrum between the two buckets. Google’s Knowledge Panel help (notably a consumer-product help center, not Search Central — which itself tells you Google doesn’t treat it as a webmaster feature) describes them as information boxes about entities in the Knowledge GraphThe Knowledge Graph is Google's database of entities — people, places, organizations, and things — and the factual relationships between them. It's separate from any single website's structured data: your schema markup is one of many possible inputs to the graph, not the graph itself. that are automatically generated from sources across the web. You can claim a panel about your own organization or person for suggested-edit access — but you can’t create one, and markup alone doesn’t produce one. Organization/Person schema with sameAs feeds the Knowledge Graph; it doesn’t guarantee a panel.

Rich results are the one genuinely markup-driven bucket — covered in full on the Rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show. page — and even they carry the “no guarantee” caveat above.

Bing’s featured-snippet / Q&A system works the same algorithmic way, which is useful confirmation that this isn’t a Google quirk. Bing’s Ali Alvi (Principal Lead PM for AI Products) and Fabrice Canel (Principal PM), interviewed by Search Engine Journal, describe Q&A as essentially the same core ranking algorithm as the blue links with different weighting, built “pretty much end-to-end” on neural networks and machine- learning models with hundreds of millions of parameters. Bing can even surface a Q&A answer from a passage that isn’t currently ranking in the traditional results, because the Q&A system keeps its own memory of candidate passages. Human annotation assists the extraction of relevant blocks — it isn’t a rankable structured-data signal you can supply. Same split as Google: neural relevance extraction on top of ranking, not a schema switch. On the opt-out side, Bing mirrors Google too — Fabrice Canel’s 2020 snippet-controls announcement gave webmasters robots meta controls over preview length: an opt-out/limit lever, not an opt-in one.

CTR impact — the real reason to care

Since SERP features don’t move rankings, their entire practical value is click- through rate, and here the data is genuinely useful.

The baseline first. Ahrefs’ study of clicks by position (last updated October 2025, using aggregated Search ConsoleA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. click/impression data through August 2025) foundA 302 (\"Found\") is a temporary redirect: it forwards users to a new URL while telling search engines the original URL should stay in the index. It's a weak canonicalization signal, not the zero-equity dead end of SEO folklore. 96.98% of desktop clicks and 97.56% of mobile clicks land in the top 10 results — the study no longer publishes a separate top-3-only breakdown, so don’t repeat an older “top 3 take X%” figure you may see elsewhere. That’s the overwhelming majority of the pie SERP features carve up.

Sistrix has the best per-feature CTR dataset I’ve seen — a mobile-SERP analysis of over 80 million keywords, first published in 2020 and last updated July 2025 — CTRs by type of search result:

Sistrix’s own summary of the whole dataset: more features = a lower CTR. Every extra feature on a SERP fragments clicks further away from position 1. That’s the redistribution effect in one line.

The flip side is that a rich result can lift your CTR relative to a plain listing — which is where the case studies from my Ahrefs structured-data guide come in, the same ones cited on the parent hub: Google’s own numbers show Rotten Tomatoes at 25% higher CTR, Nestlé at 82% higher CTR, and Food Network at a 35% increase in visits on rich-result pages. So both things are true at once: rich results can win you a bigger slice, even as SERP features overall shrink the total organic click pool.

On prevalence, a couple of numbers Backlinko cites (in its SERP-features page, last updated April 2026) are good myth-busters — worth noting these are Backlinko’s citations of other trackers, not its own study. Featured snippets get outsized attention in SEO content but appear on only about 0.24% of searches, per Semrush Sensor data from March 2026 — far rarer than people assume. People Also Ask, by contrast, shows on roughly 53% of US queries, per Advanced Web Ranking (no date given for that figure). Chase the common one, not the shiny rare one.

How the SERP has shrunk (and grown)

The structured-data side of the feature list is contracting. Reusing the exact dates established on the Structured DataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. hub:

• HowTo rich results were removed from desktop in 2023 and limited thereafter.
• FAQ rich results were restricted in 2023 to government and health sites, then fully removed by May 2026. The FAQPage type is still valid markup — it just no longer produces a SERP enhancement for most sites.
• Six more types were removed in June 2025: Claim Review, Estimated Salary, Learning Video, Special Announcement, Vehicle Listing, and the old single-course format. Book ActionsBook schema (schema.org/Book) is structured data describing a book's title, author, ISBN, and format. It can power Google's review-snippet rich result and — for approved feed providers — the registration-gated Book Actions program, and it helps Google disambiguate book entities even without a visible rich result. was announced in the same batch but was not actually removed.

Meanwhile the algorithmic side is expanding, above all AI Overviews. Ahrefs’ AI Overviews study (300,000 keywords, desktop, comparing US Search ConsoleGoogle's free tool for monitoring crawling, indexing, and search performance. data from March 2024 to March 2025 around the AI Overviews rollout) found they reduce clicks to the top-ranking organic result by 34.5%; other studies put the reduction even higher, and figures vary by device, country, and query type — treat any single number as directional, not universal. Net effect: more of the SERP is Google-generated content you can’t influence with markup, and less of it is site-controllable rich results.

The zero-click debate — both sides

This is the strategic backdrop, and it’s an open dispute, not a settled number, so I’ll give you both sides rather than pick a winner.

SparkToro’s Rand Fishkin argues the SERP is becoming a walled garden. His 2026 analysis reports that in the first four months of 2026, 68.01% of Google searches ended without a click — up from 60.45% in 2024 and roughly 45% a decade ago — and that AI Overviews, when present, reduce CTR by nearly 60% and now appear on 20%+ of searches.

Google disputes the framing. Search Liaison Danny Sullivan, quoted in Search Engine Journal’s coverage, called the zero-click narrative a “misleading claim” that “relies on flawed methodology that misunderstands how people use Search,” adding that Google sends “billions of visits to websites every day, and the traffic we’ve sent to the open web has increased every year.” Google’s specific objections: SparkToro’s method counts a broad query later refined into a narrower one as “zero-click” even when the user eventually visits a site, and it doesn’t account for navigation to apps rather than the open web. Notably, Google offers methodology objections but no comparable published figure of its own.

My read: don’t wait for this to resolve. Whatever the exact percentage, the direction is clear enough that you should assume some queries will simply never send you a click, and plan around the ones that still can.

What you can actually do about it

The practical close is short, because the whole point of the structured-data-vs- algorithmic split is that most of the SERP isn’t a lever:

1. Implement eligible structured data accurately — that’s the one bucket you genuinely control, and it can lift your CTR. Do it right (see the Structured DataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. hub) and validate it.
2. Don’t chase non-markup features. There’s no tag for a featured snippet, PAA slot, knowledge panel, or AI Overview. Write clearly and comprehensively — that’s what surfaces you in those — but stop looking for a switch that doesn’t exist.
3. Measure feature-specific CTR in Search Console. Watch your click-through by query and page; that’s where feature effects actually show up in your data.
4. Accept that some features are simply not influenceable — and don’t build a strategy around a rich result that’s already been removed.

Where to go next

This article nests under the Structured DataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. hub. The closely related deep dives:

• Rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show. — the one genuinely markup-driven bucket of SERP features: which types still exist, eligibility, the removals, and tracking them in Search Console.
• Schema markupSchema markup is code that uses the schema.org vocabulary to label what your content means so search engines can understand it and show rich results. It's most often written in JSON-LD, and it's not a direct ranking factor. — the schema.org vocabulary that unlocks rich-result eligibility.
• JSON-LDJSON-LD (JavaScript Object Notation for Linked Data) is a script-based structured data format, typically paired with the schema.org vocabulary to describe page content for search engines and AI systems. Google recommends it over Microdata and RDFa because it's the easiest format to implement and maintain at scale — but all three work, and structured data isn't a ranking signal. — the recommended format for writing that markup.

This sub-cluster lives in the broader on-page SEO cluster.

AI summary

A condensed take on the Advanced version:

• SERP featuresSERP features are any element on a search results page beyond the classic ten blue links — featured snippets, People Also Ask, knowledge panels, sitelinks, image and video packs, AI Overviews, and structured-data-driven rich results. Google documents that losing rich-result eligibility doesn't affect ranking; their SEO relevance is CTR (they redistribute clicks). Some are unlocked by markup; most are purely algorithmic. = everything beyond the ten blue links (featured snippets, PAA, knowledge panelsThe Knowledge Graph is Google's database of entities — people, places, organizations, and things — and the factual relationships between them. It's separate from any single website's structured data: your schema markup is one of many possible inputs to the graph, not the graph itself., sitelinksSitelinks are extra links from the same domain that Google clusters together under a single search result, usually for branded or navigational queries. They're generated entirely algorithmically — there's no way to add, edit, or guarantee them., image/video packs, AI OverviewsAI Overviews are the AI-generated summary box Google shows above or within its regular search results, written by Gemini models from pages retrieved out of Google's normal Search index. It's a Search feature, not a separate platform or index., rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show.). “SERP feature” is industry vocabulary; Google says “search result features” / “visual elements” and uses five buckets (text, rich, image, video, exploration).
• The spine — two mechanisms: (1) structured-data-driven rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show. (Product, Recipe, Event, Review — markup makes you eligible, never guaranteed) and (2) purely algorithmic features (featured snippets, PAA, knowledge panels, AI Overviews, image/video packs, sitelinksSitelinks are extra links from the same domain that Google clusters together under a single search result, usually for branded or navigational queries. They're generated entirely algorithmically — there's no way to add, edit, or guarantee them. — no markup lever exists). Conflating the two is the most common competitor error.
• Not a ranking factor — documented for structured dataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding.. Google’s sd-policies page: losing rich-result eligibility “doesn’t affect how the page ranks in Google web search.” That’s the one case Google has actually documented; the same eligibility-vs-ranking split is understood (not separately documented) to apply to other features too.
• Featured snippets are algorithmic: Google says you “can’t” request one; the only lever is opting out (nosnippet / max-snippet). PAA has no official Google doc — all mechanics claims are inference. Knowledge panels are Knowledge-Graph generated — claimable, not creatable.
• Bing mirrors this: its Q&A/featured-snippet system is neural-network relevance extraction on top of ranking (Alvi/Canel), not a schema switch.
• CTR is the real story. Baseline: 96.98% of desktop / 97.56% of mobile clicks land in the top 10 (Ahrefs, Oct 2025 update). Per-feature CTR (Sistrix, mobile-only, 80M+ keywords); “more features = lower CTR.” Rich results can lift your CTR (Rotten Tomatoes 25%, Nestlé 82%, Food Network 35%). Featured snippets appear on just ~0.24% of searches (Semrush Sensor, Mar 2026); PAA on ~53% of US queries (Advanced Web Ranking, undated) — both figures via Backlinko.
• The SERP is shifting: structured-data side shrinking (FAQ, HowTo, six types June 2025); algorithmic side growing (AI OverviewsAI Overviews are the AI-generated summary box Google shows above or within its regular search results, written by Gemini models from pages retrieved out of Google's normal Search index. It's a Search feature, not a separate platform or index. cut top-organic clicks ~34.5%+).
• Zero-click is contested: SparkToro says 68% of 2026 searches end without a click; Google (Sullivan) disputes the methodology without a counter-figure. Present both.
• What to do: implement eligible structured data accurately, don’t chase non-markup features, measure feature CTR in Search ConsoleA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results., accept some features aren’t influenceable.

Official documentation

Primary-source documentation from the search engines.

Google

• Visual Elements Gallery — Google’s actual taxonomy: text / rich / image / video / exploration features.
• Structured Data Search Gallery — the canonical current list of structured-data-driven rich result types.
• Featured Snippets — the “you can’t request one” statement and the nosnippet / max-snippet opt-out controls.
• Structured Data General Guidelines / Policies — the documented “doesn’t affect how the page ranks” line and the “no guarantee” line.
• About knowledge panels — Knowledge Panel Help — definition, automated generation, and claiming (a consumer-product help center, not Search Central).

Bing / Microsoft

• Announcing new options for webmasters to control their snippets at Bing — Fabrice Canel on snippet-preview controls (the opt-out lever).

Quotes from the source

On-the-record statements from Google and Bing. Where a page exposes the text, the link is a deep link that jumps to the quoted passage.

Google — the taxonomy (Visual Elements Gallery)

• “Visual elements are the building blocks of the Google Search results page that a user can perceive or interact with.” Jump to quote
• On rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show.: “A result that typically relies on structured dataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. in the markup of your page to display graphical elements or interactive experiences.” Jump to quote
• On exploration features: “A feature that helps searchers expand and refine their initial search.” Jump to quote

Google — featured snippets are algorithmic, not markup-driven

• “Special boxes where the format of a regular search result is reversed, showing the descriptive snippet first.” Jump to quote
• On whether you can request one: “You can’t. Google systems determine whether a page would make a good featured snippet for a user’s search request, and if so, elevates it.” Jump to quote

Google — not a ranking factor, and never guaranteed

• “A structured data manual action means that a page loses eligibility for appearance as a rich result; it doesn’t affect how the page ranks in Google web search.” Jump to quote
• “Google does not guarantee that your structured data will show up in search results, even if your page is marked up correctly.” Jump to quote
• “Google uses structured data to understand the content on the page and show that content in a richer appearance in search results, which is called a rich result.” Jump to quote

Bing — the Q&A / featured-snippet system is neural, not markup

Ali Alvi and Fabrice Canel described Bing’s Q&A as essentially the blue-link ranking algorithm with different weighting — “pretty much end-to-end” neural networks and machine-learning models — able to surface passages that aren’t currently ranking, with human annotation assisting extraction rather than acting as a rankable markup signal. Reported via Search Engine Journal’s interview coverage rather than a first-party Bing doc; treated here as paraphrase, not verbatim quotes.

The zero-click debate — both sides

• Danny Sullivan (Google Search Liaison): the zero-click narrative is a “misleading claim” that “relies on flawed methodology that misunderstands how people use Search.” Coverage
• Rand Fishkin (SparkToro): “In the first four months of 2026, a whopping 68.01% of Google searches ended without a click.” Source

SERP features — cheat sheet

How each feature gets selected, by feature

Three separate questions, not one yes/no: who selects it, does markup create eligibility, and what lever (if any) does a publisher actually have?

Fast facts

• Not a ranking factor — eligibility “doesn’t affect how the page ranks” (Google, sd-policies).
• Featured snippets: ~0.24% of searches, per Semrush Sensor, Mar 2026 (rare). PAA: ~53% of US queries, per Advanced Web Ranking, undated (common).
• CTR baseline: 96.98% of desktop clicks / 97.56% of mobile clicks land in the top 10 (Ahrefs, Oct 2025 update, Aug 2025 GSCA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. data).
• “More features = lower CTR” — Sistrix (mobile SERPs, 80M+ keywords, updated Jul 2025).
• Rich-result CTR wins: Rotten Tomatoes 25%, Nestlé 82%, Food Network 35%.
• AI OverviewsAI Overviews are the AI-generated summary box Google shows above or within its regular search results, written by Gemini models from pages retrieved out of Google's normal Search index. It's a Search feature, not a separate platform or index. cut top-ranking organic clicks by ~34.5% in Ahrefs’ desktop US study (300k keywords, Mar 2024–Mar 2025) and are expanding; other studies report higher.
• Removed on the markup side: HowTo (desktop, 2023); FAQ (restricted 2023, fully removed by May 2026 — FAQPage type still valid); six types June 2025 (Book ActionsBook schema (schema.org/Book) is structured data describing a book's title, author, ISBN, and format. It can power Google's review-snippet rich result and — for approved feed providers — the registration-gated Book Actions program, and it helps Google disambiguate book entities even without a visible rich result. announced but NOT removed).
• “SERP feature” is industry vocabulary; Google says “search result features” / “visual elements.”

SERP feature mistakes worth retiring

Treating “SERP feature” as one bucket with one fix. Most competing articles list featured snippets, PAA, knowledge panelsThe Knowledge Graph is Google's database of entities — people, places, organizations, and things — and the factual relationships between them. It's separate from any single website's structured data: your schema markup is one of many possible inputs to the graph, not the graph itself., and rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show. side by side with the same “add structured dataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding.” advice. It’s wrong for most of the list — only rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show. (Product, Recipe, Event, Review, Breadcrumb, and similar) are markup-driven. Instead: sort any feature you’re chasing into “structured-data-driven” or “purely algorithmic” before deciding whether markup can do anything at all.

Trying to “get” a featured snippet with markup or a special tag. Google is explicit that you can’t request one — its systems elevate a page algorithmically. Instead: write clearly and comprehensively for the query, and if you need to control snippet appearance at all, use the one real lever — nosnippet or a lower max-snippet value to opt out, not in.

Citing confident mechanics for how People Also Ask works. There’s no dedicated Google Search Central page for PAA, despite it appearing on roughly half of US queries, so any specific “how PAA is triggered” claim is industry inference, not documented fact. Instead: treat PAA mechanics as observational, and lean on documented sources (like the featured-snippet and structured-data policy pages) wherever they exist.

Assuming a rich result or featured snippet will move your ranking. Google’s own structured-data policies page says losing rich-result eligibility “doesn’t affect how the page ranks” — eligibility and ranking are separate systems. Instead: judge structured data and SERP-feature work on CTR impact, not position, and don’t promise a ranking lift from a rich result to a client or stakeholder.

Building a strategy around a rich result type that’s already gone. HowTo was removed from desktop in 2023, FAQ was fully removed by May 2026, and six more types were cut in June 2025. Continuing to implement or report on a removed type wastes effort. Instead: check the current Structured Data Search Gallery before investing in any specific rich-result type, and confirm what markup is even displaying anything for your site right now.

Chasing a knowledge panel with schema alone. Organization/Person markup with sameAs feeds the Knowledge GraphThe Knowledge Graph is Google's database of entities — people, places, organizations, and things — and the factual relationships between them. It's separate from any single website's structured data: your schema markup is one of many possible inputs to the graph, not the graph itself., but it doesn’t create a panel — panels are Knowledge-Graph generated and only claimable after the fact, never built to order. Instead: treat entity markup as a contribution to Google’s understanding, not a guaranteed panel-generation trigger.

How to measure SERP-feature impact

SERP featuresSERP features are any element on a search results page beyond the classic ten blue links — featured snippets, People Also Ask, knowledge panels, sitelinks, image and video packs, AI Overviews, and structured-data-driven rich results. Google documents that losing rich-result eligibility doesn't affect ranking; their SEO relevance is CTR (they redistribute clicks). Some are unlocked by markup; most are purely algorithmic. don’t move rankings, so the only defensible KPI here is click behavior — and the article’s own close names Search ConsoleA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. as the place to watch it.

One honest limitation up front: Search ConsoleGoogle's free tool for monitoring crawling, indexing, and search performance. gives you first-party observed query/page clicks, impressions, CTR, and position, plus per-type validity reporting in Enhancements for markup you’ve actually implemented — but it does not expose a clean, universal “this click came from a featured snippet” dimension for every SERP surface. For features it doesn’t break out (PAA and AI OverviewsAI Overviews are the AI-generated summary box Google shows above or within its regular search results, written by Gemini models from pages retrieved out of Google's normal Search index. It's a Search feature, not a separate platform or index. on most query types), you’re inferring from a before/after CTR delta, not reading a native attribution field. Third-party trackers (Sistrix, Ahrefs, Semrush and similar) fill that gap with their own dated, tool-specific SERP observations — treat those as observed-and-dated snapshots too, not a causal measurement of what caused a click change, and always note which tool, database, device, and date range a figure came from before comparing it to another tool’s number.

CTR by query, in Search Console Performance

What it tells you: whether a query that triggers a SERP feature (a featured snippet, PAA, an AI Overview, a rich result) is pulling more or fewer clicks than a comparable query without one. How to pull it: Search Console → Performance → filter/segment by the specific query, compare CTR before and after the feature appears (or against similar non-feature queries). Benchmark / realistic range: no universal number applies — Sistrix’s published ranges (e.g., ~23% pos-1 CTR with a featured snippet present vs. ~46.9% with sitelinksSitelinks are extra links from the same domain that Google clusters together under a single search result, usually for branded or navigational queries. They're generated entirely algorithmically — there's no way to add, edit, or guarantee them.) are industry-wide averages, not a target for any one site. Use your own query’s CTR history as the baseline and watch the delta, not an absolute figure. Cadence: monthly, or immediately after you notice a new feature appear on a tracked query.

CTR by page, in Search Console Performance

What it tells you: whether a specific page’s overall click-through changed after you added or fixed structured dataStructured data is a standardized way of labeling page content (using the schema.org vocabulary in JSON-LD, Microdata, or RDFa) so search engines can understand its meaning. It's not a direct ranking factor — its value is rich results and entity understanding. (the one lever you actually control). How to pull it: Search Console → Performance → Pages, compare the page’s CTR trend for several weeks before vs. after the markup change; cross-check against impressions to rule out a ranking-position shift as the real cause. Benchmark / realistic range: depends entirely on the page’s prior CTR and its rich-result type — set your own pre-change baseline rather than assuming an industry figure like Nestlé’s 82% or Rotten Tomatoes’ 25% applies to your page. Cadence: each time you ship or edit structured data on a page, then monthly thereafter.

Rich-result eligibility and errors, in Search Console Enhancements

What it tells you: whether your structured data is actually valid and eligible for a rich result at all — a precondition for any CTR effect, since Google never guarantees display even on correct markup. How to pull it: Search Console → Enhancements (per rich-result type), or spot-check any URL with the Rich Results Test. Benchmark / realistic range: the only honest target is zero “Invalid” items reported for the type you’re implementing — there’s no meaningful industry average for eligibility rates across sites. Cadence: check after every markup deployment, then monthly for regressions.

Impressions share for feature-eligible queries

What it tells you: whether the queries you’d expect to trigger PAA, an AI Overview, or a featured snippet are shrinking your organic impressions over time, which is the leading indicator behind the zero-click debate. How to pull it: Search Console → Performance, track impressions (not just clicks) for the query set you care about across quarters. Benchmark / realistic range: no defensible external number — the zero-click studies cited in this article disagree with each other and with Google’s own framing, so establish your own quarter-over-quarter trend line instead of adopting SparkToro’s or anyone else’s percentage as your target. Cadence: quarterly.

Test yourself: SERP Features

Five quick questions on what SERP featuresSERP features are any element on a search results page beyond the classic ten blue links — featured snippets, People Also Ask, knowledge panels, sitelinks, image and video packs, AI Overviews, and structured-data-driven rich results. Google documents that losing rich-result eligibility doesn't affect ranking; their SEO relevance is CTR (they redistribute clicks). Some are unlocked by markup; most are purely algorithmic. are, which ones markup unlocks, and how they affect CTR. Pick an answer for each, then check.

SERP Features

SERP features are any element on a search results page beyond the classic ten blue links — featured snippets, People Also Ask, knowledge panels, sitelinks, image and video packs, AI Overviews, and structured-data-driven rich results. Google documents that losing rich-result eligibility doesn't affect ranking; their SEO relevance is CTR (they redistribute clicks). Some are unlocked by markup; most are purely algorithmic.

Related: Rich Results, Structured Data

SERP Features

SERP features (search engine results page features) are any element on a results page beyond the classic “10 blue links” — featured snippets, People Also Ask, knowledge panelsThe Knowledge Graph is Google's database of entities — people, places, organizations, and things — and the factual relationships between them. It's separate from any single website's structured data: your schema markup is one of many possible inputs to the graph, not the graph itself., local packs, sitelinksSitelinks are extra links from the same domain that Google clusters together under a single search result, usually for branded or navigational queries. They're generated entirely algorithmically — there's no way to add, edit, or guarantee them., image and video packs, review stars, AI OverviewsAI Overviews are the AI-generated summary box Google shows above or within its regular search results, written by Gemini models from pages retrieved out of Google's normal Search index. It's a Search feature, not a separate platform or index., and the dozens of structured-data-driven rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show. (Product, Recipe, Event, Review, and so on). They change how a result is displayed, not whether or where it ranks.

“SERP feature” is industry and tool vocabulary (Ahrefs, Semrush, Sistrix). Google’s own documentation doesn’t use the term — it calls them “search result features” or “visual elements,” and groups them into five buckets: text result, rich result, image result, video result, and exploration features.

The single most useful distinction is how each feature is triggered:

• Structured-data-driven rich resultsRich results (formerly 'rich snippets') are enhanced search listings — stars, images, prices, breadcrumbs, video thumbnails, and more — that Google and Bing build from structured data. They're a display feature, not a ranking factor, and eligibility never guarantees they'll show. are unlocked by adding schema.orgSchema markup is code that uses the schema.org vocabulary to label what your content means so search engines can understand it and show rich results. It's most often written in JSON-LD, and it's not a direct ranking factor. markup to a page — but eligibility is never guaranteed even with correct markup.
• Purely algorithmic features (featured snippets, People Also Ask, knowledge panels, AI OverviewsAI Overviews are the AI-generated summary box Google shows above or within its regular search results, written by Gemini models from pages retrieved out of Google's normal Search index. It's a Search feature, not a separate platform or index., image/video packs, sitelinksSitelinks are extra links from the same domain that Google clusters together under a single search result, usually for branded or navigational queries. They're generated entirely algorithmically — there's no way to add, edit, or guarantee them.) have no markup lever at all. Google states outright that you can’t request a featured snippet.

Two more points trip people up: Google explicitly documents that losing structured-data eligibility “doesn’t affect how the page ranks in Google web search” — that’s the one case Google has actually put in writing, and the same eligibility-vs-ranking split is understood (not separately documented) to apply to the rest — and their real SEO impact is on click-through rate, which they redistribute, often away from the traditional organic listing they sit beside.

Related: Rich Results, Structured Data

Changelog