What is Featured Snippet?

A featured snippet, sometimes called an answer box or "position zero", is a block Google shows above the normal ranked list, pulling a specific excerpt, a paragraph, a numbered list or a table, directly from one webpage it judges best answers the query. Unlike an ordinary organic listing, which shows a title and meta description, a snippet shows the answer itself with a link crediting the source. The defining property is that it is extraction from a single page: the displayed text exists word for word on that page.

Snippets are selected algorithmically from pages Google already ranks well for the query, and they favour content that answers plainly and early. A direct one or two sentence answer to a "what is" question, a clean numbered list for a "how to", a well-structured table for a comparison. Content buried under several paragraphs of preamble rarely wins the snippet even when the page ranks well overall.

The important thing to know in 2026 is that this feature is no longer where most of the attention should go. As AI Overviews expanded, featured snippets declined sharply, and the two now rarely appear together. On English-language informational queries in the US, the position a snippet used to occupy is usually filled by an Overview instead. Every dataset agrees on that direction. Published estimates of the size of the drop disagree with each other substantially, so treat any specific percentage you see with caution unless it comes with a sample size, date and country.

Snippets have not disappeared. They still turn up on short factual lookups, on some step-by-step content, and in markets and languages where AI Overviews have not fully rolled out. If those describe your audience, the feature is still worth having.

The more durable point is that the structural work behind a featured snippet did not become obsolete when the feature faded. Answering directly, early, in a self-contained passage is what made a page snippet-eligible, and it is also what makes a passage easy for a generative system to lift and cite. Optimising for snippets was, in retrospect, a decade-long rehearsal for AEO.

One distinction is worth making explicitly, because most advice collapses it. What Google ended was a search feature, not the underlying data. FAQ rich results ended on 7 May 2026, and the Search Console report and Rich Results Test support are reported to follow in June 2026, with API support in August 2026. So FAQPage markup no longer earns a Google SERP treatment, and by those reports the measurement tooling is going too.

That is not the same as saying marked-up question-and-answer pairs are worthless to everything else reading the page. The mechanism is reasonable: explicit markup makes the question-to-answer boundary unambiguous where prose leaves a parser to infer it. Plausible, not proven. A July 2026 survey of 45 studies found no reviewed technique with a stable, cross-platform causal effect on discoverability, and this belongs in that same category. Treat it as cheap and possibly useful to an AI crawler, not as a lever with a known return, and do not spend effort stripping markup you already have.

Keep llms.txt out of this argument. It is a different kind of thing, an access convention aimed at AI crawlers rather than a search feature, so it has no deprecation to reason around. Google says llms.txt is not needed for Google Search and has no effect on visibility or rankings (Google, June 2026). Other AI systems may read it, but support there is unconfirmed.