AI Summary
Featured snippets and AI Overviews are not the same feature, do not select content the same way, and do not deliver the same traffic outcomes. Understanding the difference determines which optimization tactic to prioritize in 2026.
Ahrefs analyzed 1,000,000 US desktop SERPs from their Keywords Explorer database comparing January 2025 to June 2025. Featured snippets appeared on 5.53% of SERPs in June, down from 15.41% in January, a 64% relative decline. AI Overviews grew from 3.93% to 27.43% over the same period, a 598% increase. The correlation between AI Overview growth and featured snippet decline is -0.90, and a clear switch-over point appears in March 2025 coinciding with a 116% spike in AI Overview prevalence. A separate Ahrefs study published in February 2026 found that AI Overviews now correlate with a 58% lower average click-through rate for the top-ranking page versus the same keyword without an AI Overview.
These two data points define the current situation: featured snippets are being displaced by AI Overviews, and AI Overviews reduce individual-page CTR while appearing on a growing share of queries. Neither feature is dead. Both require optimization. The tactics diverge significantly.
What each feature is and how selection differs
A featured snippet is a single-source extraction. Google identifies one page that directly answers a query and pulls a short passage, list, or table into a formatted box at the top of the SERP. The page must rank in the top organic positions to be eligible. The snippet reflects what is already on that page verbatim or near-verbatim. Selection criteria center on directness of answer, passage length alignment with the query type, and structured formatting.
An AI Overview is a multi-source synthesis. Google’s model reads multiple pages and generates a new summary response, with inline citations to the sources it drew from. The cited pages do not need to rank in position one. Our May 2026 citation study across 153,425 citations found that 76.95% of cited URLs are not in the organic top-10 for the query. A page can be an AI Overview source without ranking on page one at all. Selection criteria include topical authority, entity consistency, structured content, and the quality of the specific sentences that answer the query, not the overall page ranking.
The practical consequence: optimizing for a featured snippet means winning a single ranking battle. Optimizing for AI Overviews means building citation authority across a topic cluster. These are different projects requiring different content investments. Our Google AI Overviews optimization playbook covers the full citation-building approach.
What happens to snippets when an AI Overview appears
The Ahrefs SERP feature data tells the story directly. The -0.90 correlation between AI Overview growth and featured snippet decline is not coincidence. Both features resolve informational queries inline without requiring a click. When Google has sufficient confidence to generate an AI Overview for a query, the featured snippet becomes redundant. It is suppressed.
The switch-over pattern in the Ahrefs data is the clearest evidence available. In March 2025, AI Overviews grew by 116%. Featured snippets dropped in parallel. The relationship is structural: Google is replacing a one-source extraction model with a multi-source synthesis model. Featured snippets will continue to decline as AI Overview coverage expands to more query types.
The queries where featured snippets still reliably appear are those where AI Overviews have not yet displaced them: highly factual lookups (definitions, unit conversions, single-number answers), some navigational queries, and certain how-to formats. For most informational SaaS and B2B queries, the AI Overview is now the primary answer surface.
Traffic dynamics: CTR and volume compared
Featured snippets historically delivered 8-12% click-through rates for the owning page on informational queries. That CTR was high because the snippet created curiosity and signaled the page as the authoritative source. Users who wanted more detail clicked through.
AI Overviews work differently. The Ahrefs February 2026 study found a 58% lower average CTR for position-one pages on queries with an AI Overview versus comparable queries without one. Our own AI Overviews CTR analysis is consistent with this range. The AI Overview resolves the query inline. Cited sources receive some traffic, but the zero-click rate is significantly higher than with featured snippets.
The net traffic calculus depends on query volume and position. A page that owned a featured snippet on a high-volume query and now sees that snippet replaced by an AI Overview will typically experience a traffic decline even if it becomes an AI Overview citation. However, AI Overview citations can appear across many more queries than a single featured snippet, and the Ahrefs data shows AI Overviews now appear on 27.43% of SERPs versus 5.53% for snippets. A citation strategy that spans many mid-volume queries can outperform a single high-CTR snippet at scale.
Bain and Company’s research adds the broader context: zero-click search is now the default for consumers, with click-through rates falling by as much as 30% in some categories, including B2B software. The direction of travel is unambiguous. Both features contribute to zero-click behavior, and AI Overviews are accelerating it.
Does snippet optimization still matter?
Yes, for two reasons. First, featured snippets still appear on 5.53% of SERPs and that is not zero. Queries where AI Overviews have not displaced snippets remain winnable and continue to deliver higher per-click CTR than AI Overview citations. Second, the content tactics that win featured snippets overlap significantly with AI Overview citation tactics. A direct, concise, well-structured answer that wins a snippet is often also a strong AI Overview citation candidate.
The overlap is not complete, though. Snippets require top-page organic ranking. AI Overview citations do not. A page that ranks position 15 can be an AI Overview source if it contains the right atomic facts. Our May 2026 study found that the mean cited sentence is 9.27 words, median 10, with no cited sentence exceeding 18 words. Snippet optimization typically targets 40-60 word passages. These are different content units requiring different writing approaches.
Practical recommendation: continue optimizing for featured snippets on queries where they still appear, but do not build an entire content strategy around snippet capture. The underlying mechanics are being superseded. For the Google AI Mode trajectory and what it means for organic content strategy, the playbook covers how AI Mode extends zero-click behavior even further.
Optimization tactics compared
The tactics diverge clearly once you understand the selection mechanisms. Featured snippets reward page-level ranking signals plus passage-level directness. AI Overviews reward entity authority plus sentence-level extractability across the document.
| Factor | Featured Snippets | AI Overviews |
|---|---|---|
| Selection mechanism | Single-source extraction from top-ranked page | Multi-source synthesis; cited pages need not rank top-10 |
| Organic ranking required | Yes, typically top-5 | No: 76.95% of cited URLs outside organic top-10 |
| Content unit | 40-60 word direct answer passage | Short atomic sentences (mean 9.27 words) |
| Ideal content position | Early in page, under matching H2 | First 50% of document (74.9% of citations in first half) |
| CTR when triggered | 8-12% historically | ~58% lower vs. non-AIO equivalent (Ahrefs Feb 2026) |
| SERP prevalence (June 2025) | 5.53% of SERPs | 27.43% of SERPs |
| Trend direction | Down 64% Jan-Jun 2025 | Up 598% Jan-Jun 2025 |
| Key optimization lever | Rank in top-5, structure concise answer under H2 | Topical authority, entity consistency, atomic sentence writing |
Featured snippet tactics (still worth doing)
- Target queries where AI Overviews are not yet present (check via manual sampling or rank tracker SERP feature data).
- Place a 40-60 word direct answer immediately under the target H2, matching the query phrasing as a question in the heading.
- Use parallel list structure for how-to and step-based queries; Google extracts numbered lists for procedural snippets.
- Add a comparison table for queries with “vs” or “difference between” phrasing.
- Maintain top-5 organic ranking through standard on-page and link signals.
AI Overview citation tactics (the growth lever)
- Write in atomic sentences of 6-15 words. Our May 2026 study found the 6-10 word range accounts for 45.2% of all cited sentences. Each sentence should contain one verifiable fact.
- Place your most citable sentences in the first half of the document. 74.9% of citations from our May 2026 study came from the first half of the source page.
- Build topical authority across a cluster. AI Overviews pull from multiple pages on related topics. A single page rarely dominates; a well-structured cluster creates citation surface area across many queries.
- Use original data and statistics. The arXiv GEO paper confirmed statistics increase AI impression scores by +30-40%. Proprietary figures are more citable than recycled industry numbers.
- Maintain entity consistency: brand name, author credentials, and topic associations should be consistent across your site, social profiles, and third-party mentions.
The atomic sentence SEO guide covers the writing mechanics in full, with examples of how to restructure existing paragraphs into citable sentence units. For a broader view of how citation authority compounds over time, the GEO overview sets the strategic context.
What this means for content investment decisions
The question we hear most often from content teams is: “should we stop chasing snippets and shift everything to AI Overviews?” The answer is no, but the resource allocation should shift substantially toward AI citation tactics. Snippets are a declining surface. AI Overviews are a growing one. A content program that still optimizes exclusively for snippets in 2026 is fighting the wrong battle.
The more productive frame is: every piece of content you produce should satisfy both signals. Short, direct answers under H2 headings satisfy snippet extraction. Those same direct answers, written in atomic sentence structure, satisfy AI Overview citation. The overlap is large enough that you rarely need to make a direct trade-off between the two.
Where you do need to make a decision is in new content investment. A page built primarily to rank and capture a snippet on a single query is a narrower bet than a page built to establish topical authority and attract AI citations across a cluster of related queries. The latter produces durable citation value even as individual snippet opportunities decline.
Our GEO audit identifies which of your existing pages are already generating AI Overview citations, which are snippet-optimized but not citation-ready, and which gaps represent the highest-value content investment. The audit costs $4,500 flat and delivers a prioritized brief for both snippet and citation optimization. See the GEO audit checklist for the 50-point evaluation methodology. For ongoing share-of-voice tracking across AI platforms, our AI citation tracking service starts at $1,500/month.