Featured snippet

Google picks a featured snippet from pages that already rank on the first page for the query. That single fact settles most of the strategy: the work is ranking, and the formatting is what decides which of the ranking pages gets lifted.

There is no tag, no attribute and no schema type that requests one. Anybody selling snippet markup is selling something that does not exist, and the FAQ markup often confused with it is a different feature that Google cut back in 2023.

What does influence selection is whether the page contains a clean, self-contained answer that matches the shape of the question. A definition query wants a short paragraph. A how-to query wants an ordered list. A comparison query wants a table. Google is looking for something it can lift whole, and a page that buries the answer in the middle of a paragraph gives it nothing liftable.

Position on the first page matters less than people assume. A result at five with a clean answer regularly takes the snippet over a result at one that never states the thing plainly.

The practical technique is unglamorous and works. Put the question as a heading, phrased the way somebody would type it, and answer it immediately underneath in roughly forty to sixty words, as a complete statement that makes sense with no context around it.

Complete is the operative word. An answer beginning "it depends on several factors" or "as we discussed above" cannot be lifted, because lifted out of the page it says nothing. The sentence has to survive being read on its own, which is the same test that makes it useful to a reader skimming.

Match the format to the question type rather than defaulting to prose. Steps belong in an ordered list, because that is what gets displayed for a how-to. A comparison of three plans belongs in a table with a header row. Putting a list of steps into a paragraph is the most common reason a page that should own a snippet does not.

The rest of the page still has to earn its ranking. A page consisting only of a forty word answer will not rank for anything, and the snippet is selected from pages that rank. The answer goes near the top; the depth that earns the position goes below it.

A featured snippet is prominence and it is also a risk, because an answer displayed completely can satisfy the searcher without a click. Both things are true and which dominates depends entirely on the query.

For a factual query with a short answer, the snippet often ends the visit. Somebody who wanted to know what a term means has what they wanted. That is not necessarily a loss, since the brand was the one that answered, but it should not be counted as traffic.

For a query where the answer opens a decision, the snippet works the other way. A definition that makes somebody realise their agency has been measuring the wrong thing produces a click, because the next question is harder than the one displayed. Commercial and comparison queries behave the same way.

The useful discipline is to decide which kind a page is before optimising for the snippet. Some pages exist to be read and some exist to be found, and winning a snippet on a page that needed the click is a result that looks good in a report and worse in the analytics.

Generated summaries now sit above the results for many informational queries and do the same job a featured snippet did, on a larger scale and drawing from several pages at once. On those queries the featured snippet is often no longer the top thing on the page, and sometimes is not shown at all.

The technique survived the change better than the surface did. The content that gets cited in a generated summary looks like the content that used to win snippets: a plain statement of the answer, near a heading that matches the question, in a page that has some standing behind it.

What changed is the measurement. A citation in an overview produces far less traffic than a snippet did, and it does not appear as a position in any ranking tool. A page can be the source an answer was drawn from and show nothing in a report, which is a difficult conversation with a client who is looking at sessions.

The practical response is to stop treating snippet ownership as a metric and treat it as a side effect of writing clearly. It was always fragile, it was never contractible, and it is now less measurable than it was.