Featured Snippets: -64% Visibility, 11.4% CTR When They Appear

Why featured snippets are losing importance in 2026 but still bring clicks

For years, featured snippets were the most coveted organic SERP feature: the highlighted answer box above the classic results, often referred to internally as "position zero." In 2025, their role shifted drastically. Based on one million US desktop SERPs, the feature appeared on only 5.53% of SERPs in June 2025, compared with 15.41% in January 2025, a relative decline of 64%. In the same analysis, this drop correlates very strongly (-0.9053) with the growth of AI Overviews. The data comes from the US market; the methodology is based on randomly selected desktop SERPs.

Where the feature is still shown, it continues to bring clicks, though far fewer than before: according to First Page Sage, the snippet CTR is 11.4% in 2026, while the 2025 report still showed 42.9%. For B2B companies in the DACH region, this creates two clear jobs: use snippet optimization deliberately where the format still appears at all, and use the same extractable answer structure to stay visible in AI Overviews and People Also Ask. For topics that need explaining and have long buying cycles, this on-page discipline decides brand visibility in the answer phase.

Which snippet formats are still shown in 2026

Google still shows featured snippets in four basic formats, but their frequency has shifted:

Paragraph snippets: short prose answers to definition and "what is" questions. They are the most common format and exactly the type that AI Overviews increasingly displace for purely informational queries.

List snippets: numbered or unordered lists for processes, rankings, and enumerations. They survive more often for how-to and comparison questions, because a clear sequence of steps is hard to replace with prose.

Table snippets: structured data such as prices, specifications, or comparison values. For B2B products with detailed specs, they are a stable anchor, provided the table is cleanly marked up in HTML.

Video snippets: highlighted video clips with timestamps, usually from YouTube. Relevant for demonstration-heavy topics, but rare in classic text-based B2B.

The strategic point: AI Overviews mainly take over informational top-of-funnel queries where a single paragraph delivers the answer. Factual, transactional, and ambiguous queries keep classic snippets more often. For on-page practice, this means you keep maintaining snippet structures but stay realistic about what pure definition snippets can deliver.

Visibility and click-through rate: what was measured in 2025 and 2026

Both developments can be quantified. The probability that a snippet appears for a query at all has dropped: from 15.41% in January to 5.53% in June 2025. Where the snippet appeared, its CTR in the 2025 report was 42.9%, compared with 39.8% for the organic top result. In the September 22, 2026 update, click-through rates collapsed across the board: the snippet reaches 11.4%, position 1 without an AI answer 22.6%, position 1 with an AI answer 3.6%, and 7.1% on average. Against position 1 without an AI answer, the snippet has lost its lead; against the position 1 average, it is still ahead.

Historical context helps keep expectations in check. An older, broad study showed a different picture: back then, a snippet in position 1 received only about 8.6% of clicks, while the page ranking directly below it reached about 19.6%. Without a snippet, that page received about 26% of clicks instead. The two studies measure different points in time with different methods; together they show that snippet CTR depends heavily on SERP layout and query type. In the same older analysis, about 12.29% of search queries had a featured snippet, a share that had dropped significantly by 2025.

This leads to a pragmatic prioritization. On-page optimization for snippets pays off especially for bottom-of-funnel searches, factual or transactional queries, and brand or product terms, where a click has real business value. It is less worthwhile for broad informational questions that an AI Overview answers anyway and that rarely trigger a classic snippet anymore.

On-page mechanics for position zero

The structure that wins snippets is the same one AI systems prefer to cite: precise, extractable answers.

Question-and-answer architecture: Phrase H2 or H3 headings as the question the user asks, and deliver a compact, complete answer directly below. The answer block should make sense without the context of the rest of the page.

Lead with definition sentences: Start the answer block with a clear declarative sentence following the pattern "X is ...". Google prefers to extract definitions from constructions like this.

Clean heading hierarchy: A logical H1-H2-H3 outline helps crawlers and AI systems match answer blocks to their questions. Avoid skipped levels and purely decorative headings.

Correctly marked-up lists and tables: Use real HTML (<ol>, <ul>, <table>) instead of imitation layouts built with line breaks or images. Only structured markup can be extracted reliably.

Answer before depth: Give the short answer first, then the detailed explanation. This "answer first" logic serves snippet extraction and impatient readers at the same time.

These mechanics are good information architecture: they make content easier to understand for people and machines.

On-page schema markup after the end of FAQ and HowTo rich results

The structured data that used to trigger high-reach rich results has largely lost its value. Google announced that FAQ rich results would only be shown for well-known, authoritative government and health websites and would no longer appear regularly for all other sites. HowTo rich results were switched off as of September 13, 2023 and have been considered deprecated since then. Important in practice: unused markup does no harm, but according to Google it also has no visible effect in Search.

This creates a clear separation. Markup types with active rich results, such as Product and Article, affect the SERP and therefore remain a priority. Since January 2025, Google shows breadcrumbs only in desktop results; on mobile, only the domain appears. Organization markup influences which logo Google shows in search results and in the knowledge panel. For most B2B sites, by contrast, FAQPage and HowTo markup now serve only page understanding and potentially GEO visibility, no longer the classic SERP display.

JSON-LD instead of inline markup: Implement structured data as JSON-LD in the <head> or <body>. It is Google's recommended format and the easiest to maintain.

Consistency with visible content: Markup must reflect exactly the content that users can see. Discrepancies risk structured data violations.

Organization and Author establish entities: Clean Organization and Person markup anchors your brand and authors as recognizable entities, which increasingly matters for E-E-A-T and AI source selection.

People Also Ask: the more frequent and more stable feature

While classic snippets are becoming rarer, People Also Ask remains by far the more widespread SERP feature. In a 2020 Semrush study of one million US keywords, PAA appeared on 49.37% of desktop SERPs and 52.27% of mobile SERPs. In this study, PAA was therefore about six times more likely than a featured snippet. Since the 2025 snippet decline, the gap has probably grown larger. The paragraph format also dominated in PAA: 79% of PAA boxes were paragraphs from website content. The data comes from the US and does not transfer directly to the DACH market, but it shows the order of magnitude.

For on-page work, this leads to a concrete approach:

Research PAA clusters: Collect the PAA questions shown for your main keyword and group them by topic. They reveal how your target audience actually asks.

Answer in question-and-answer blocks: For each relevant PAA question, build a separate question-and-answer block with a compact answer followed by more depth.

Link answer hubs internally: Connect these blocks with each other and with central topic pages where it makes sense, so that crawlers recognize the answer hubs as a coherent cluster.

Because PAA is more stable and more frequent than the classic snippet, it is the more productive on-page lever for many B2B sites.

Featured snippets and AI Overviews: the same bridge

The central strategic insight: snippet optimization and GEO are technically related. Both reward precise answer blocks that make sense on their own. If you write for position zero, you are also writing for citability in AI answers.

The click dynamics have shifted, though. A behavioral study showed that users who saw an AI summary clicked a classic result in 8% of visits, compared with 15% without a summary. A link within the AI summary itself was clicked in only 1% of visits. The session also ended in 26% of cases with AI summaries, compared with 16% for purely classic results. These figures come from a US study with 900 people and a tracking dataset; they illustrate the zero-click risk, but they do not represent DACH figures.

A recent re-run of a click study adds to this: the presence of an AI Overview now correlates with a 58% lower average click-through rate for position 1 content. It compared December 2023 with December 2025. The predecessor study from April 2025, covering one year instead of two, had measured 34.5% fewer clicks on top content. For DACH B2B, this means that visibility as a cited source in the answer gains weight, even as the direct click becomes rarer.

Title tags and meta descriptions as CTR levers

Where a click is still possible, the search listing decides the CTR. The levers here are measurable. An analysis of four million search results shows that title tags with 40 to 60 characters have an 8.9% better average CTR than titles outside this range. Positively phrased titles also achieve a 4.1% higher absolute CTR than negative ones. For scale: in the same study, the organic top result received 27.6% of all clicks. These values come from the international dataset.

Stick to the length sweet spot: Keep titles in the 40- to 60-character range so they are neither truncated nor look too short. Google renders by pixel width, so the character count is only an approximation.

Test question and benefit formats: Question phrasing and concrete benefit statements can raise perceived relevance, especially when they mirror your target audience's PAA language.

Descriptions as a click incentive, not a ranking factor: Meta descriptions don't influence rankings, but they shape the click. Expect Google to rewrite them often, and still provide a concise suggestion that stays close to the query.

Form of address and technical terms: Choose the form of address that fits your target audience. For relevance, what counts are the technical terms decision-makers actually search with.

E-E-A-T and helpful content as a prerequisite for selection

Snippet and AI selection require trust. Thin blocks trimmed purely for extraction are no longer enough. Since FAQ markup no longer triggers rich results for most sites, attaching FAQ blocks en masse brings no SERP advantage. What counts instead is substance and a traceable origin.

Make authors visible: Give expert articles real authors, qualifications, and Person markup. Recognizable expertise supports trust signals.

Back up sources and data: Link to primary sources and use your own data where possible. Original figures from your own practice are hard to replace and get cited preferentially.

Keep content current: Date and update content visibly. Outdated values undermine selection as a snippet or AI source.

Depth instead of a thin layer: Answer the question completely instead of with an isolated sentence. Helpful content logic rewards content that actually resolves the searcher's need.

Measure, don't guess: position zero and PAA tracking

Without targeted monitoring, visibility losses in SERP features often surface late, because rankings look stable while features disappear.

SERP feature reports: In your rank tracking, follow which keywords show snippets, PAA, and AI Overviews and whether you are the source. The market decline from 15.41% to 5.53% snippet coverage shows how quickly the mix of SERP features changes.

Position zero monitoring: Watch specifically whether snippets you have won are retained. Losses are an early indicator of a changed SERP layout.

CTR by position in GSC: Analyze CTR by position in Google Search Console. If CTR drops while the position stays stable, that points to new features above your result.

Track PAA coverage: Since PAA appeared about six times more often than snippets in the 2020 Semrush study, systematic tracking of PAA presence is especially worthwhile.

The DACH and Austrian reality

The Austrian and wider DACH market has a few peculiarities. According to StatCounter, Google holds more than 80% of the search market here. In September 2026, its share was 80.88% in Austria, 87.84% in Germany, and 80.21% in Switzerland. The Google mechanics described here therefore largely determine the market. Keep in mind that the cited studies mostly reflect US data: the orders of magnitude transfer, the absolute values do not.

German-language B2B queries follow their own patterns. Precise technical terms and local entities such as registered company names, locations, and industry terms shape search intent. B2B topics that need explaining benefit in particular from snippet and PAA optimization, because decision-makers search with detailed, often multi-part questions. If you answer these questions cleanly on the page in question-and-answer blocks, you serve classic snippets, People Also Ask, and AI Overviews with the same content architecture and secure visibility across all answer surfaces.

Going further

In 2026, featured snippets no longer take care of themselves, but they are not a lost channel either. The right response: embed snippet optimization in a broader answer strategy. That means writing extractable answer blocks that serve snippets, PAA, and AI Overviews at the same time, maintaining the schema markup that affects the SERP, and measuring feature visibility consistently. The related on-page topics are also worth a look: heading hierarchy and content structure, on-page schema markup, People Also Ask, and on-page GEO, all covered in more depth in this knowledge base.