Skip to content
Join the waitlist

Classic Search

SERP Features

SERP features are the parts of a search results page other than the standard organic links, such as AI Overviews, featured snippets, People Also Ask boxes, image and video results, knowledge panels and rich results. Which features appear depends on the query, and on the searcher’s device, country and language.

  • Industry term

By QueryRev ResearchPublished Updated6 min read

Why it matters.

Google calls them visual elements. Its gallery describes them as the building blocks of the results page: text results, rich results, image and video results, and exploration features, with the way each looks changing over time and with the device, the country and the language of the search[1]. “SERP features” is the name search teams use for the same parts.

The query decides which of them appear. Google’s systems read words such as “cooking” or “pictures” as a sign that recipes or images may suit the search best, and a search for “pizza” brings nearby businesses[2]. Google’s own example is a search for “bicycle repair shops”, which will likely show local results and no image results, where “modern bicycle” is more likely to show images but not local results[3].

So a position in a list says less than it did. A featured snippet reverses the format of a result and shows the descriptive snippet first, and it can also appear inside a related questions group[4]. Rich results are made from structured data that a page carries[5], and AI Overviews are shown only when Google’s systems judge them additive to classic Search, so they often do not appear[6]. A page can hold the first organic link and still sit below several of these.

Not every feature can be earned the same way. Google says a page cannot be marked as a featured snippet: its systems decide[4]. For the AI features of Search, it says the control for site owners is robots.txt for Googlebot, and that Google-Extended limits AI training and grounding in some of its other systems[6]. In QueryRev’s study, Google-Extended is blocked by name on 411 of the 6,294 sites whose robots.txt could be read, 6.5%[7]. That rule is not the control for the AI features of Search.

How it is measured.

Per tracked query, device and location: the features on the page in the order they appear, which of them hold a link to your site, and where the first organic link sits. Search Console numbers each element of the page as one position, from the top of the page, whether it holds one link or many[8]. A link below four elements is reported at position 5, even when it is the first organic link.

Keep the sample fixed. Because the page changes with the device, the country and the language, a feature seen on a phone in one country says little about a laptop in another. Record the same queries from the same places each month, and compare the share of queries on which each feature appears.

For the features you hold, record the feature and the URL it links. Sites that appear in Google’s AI features are reported in the Performance report of Search Console, within Web search, with the rest of the results[6]. Record the feature beside the report’s figures, so that a change in clicks can be read against a change in the page.

Measured by

  • Features per query
  • Features held
  • First organic position

A worked example.

Sample data: an invented site, to show the method.

For one sample query, a shop’s sizing guide ranks first among the organic links. The results page shows an AI Overview, a featured snippet, a People Also Ask box and a row of video results above it. In Search Console the guide reports position 5, and the click-through rate is low for a first organic result.

Nothing is wrong with the page. The features answer part of the question before the link is reached. The shop checks which of the four it could hold: the featured snippet, by putting the answer in the first sentence under the question as a heading, and the AI Overview, which needs no markup beyond the basics. It cannot choose to be in the People Also Ask box.

One results page, top to bottom

What to do, step by step.

  1. Choose the queries that matter and fix the device, the country and the language they are checked from.

  2. Snapshot each results page on a schedule and record its features in order, with the position of the first organic link.

  3. Sort the features by what can be done: a rich result needs matching structured data, a featured snippet needs a page that answers the question, an AI feature needs a page that is indexed and eligible for a snippet, and a People Also Ask box cannot be controlled.

  4. For each feature you could hold, read the page as the feature would use it: the question as a heading, the answer in the first sentence, markup that matches the visible text.

  5. Read click-through rate beside the features of the page, not beside position alone.

  6. Check the same queries again each month and compare periods, because the features of a page come and go.

Mistakes to avoid.

  • Reading position as place on the page. A position counts the elements above a link, and a feature is an element. The first organic link can be reported at position 5 while the reader has already met several answers above it.

  • Trying to mark a feature that is chosen for you. A featured snippet cannot be requested and a People Also Ask box cannot be controlled. The work is the page that answers the question well, not a tag.

  • Checking from one place and one device. The same query returns different features by device, country and language. One laptop in one office is a sample of one.

  • Planning around how a feature looks. Google says the look of its elements changes over time. A plan that depends on one feature’s appearance depends on a design Google can change.

How QueryRev treats it.

QueryRev reads impressions, clicks and position by query and page from Search Console, and keeps what the results page showed on tracked queries beside them, so a fall in clicks at a steady position is checked against a new feature before any page is rewritten. AI Overviews are one of the surfaces it tracks. See Search Intelligence and the guides to ranking in AI Overviews and tracking them.

Questions people ask.

What is the difference between a SERP feature and a rich result?

A rich result is one kind of feature: a result that typically relies on structured data to show graphical or interactive elements. Other features, such as featured snippets, People Also Ask boxes and AI Overviews, do not come from markup you add.

Can I choose which SERP features my page appears in?

Only some. Rich results need matching structured data and eligibility. A featured snippet is chosen by Google’s systems and cannot be marked. AI features need a page that is indexed and eligible for a snippet, with no extra markup.

Why does my position differ from what I see on the page?

Search Console numbers each element of a results page as one position from the top, so features above your link count before it. A link can be the first organic result and still report position 5.

Sources and citation.

Primary sources only: the search engines, the AI vendors, standards bodies and QueryRev’s own research. Each is cited in the text by its number.

  1. Automatically generating and ranking resultsGoogle Search ·

  2. In-depth guide to how Google Search worksGoogle Search Central ·

  3. AI features and your websiteGoogle Search Central ·

  4. State of AI crawler access, October 2026QueryRev Research ·

  5. What are impressions, position, and clicks?Google Search Console Help ·

Cite this definition

A suggested citation. Its link goes to the definition, whose anchor does not change.

QueryRev Research. “SERP Features.” QueryRev AI glossary, 23 October 2026. https://queryrev.com/glossary/serp-features#definition

Related terms