# Search Intent: What It Is, Types and Examples

> What search intent is, the types Google uses in its rater guidelines and which page of your store answers each query.

URL: https://clione.ai/glossary/search-intent/
Lang: en

Two people type "travel tripod" into Google. One wants to know what to look at before choosing one; the other has already decided and wants to buy it today. The words are the same and **what they are after is not**. That is search intent: the goal behind a query. It matters to a store because it decides which of your pages has a chance at each query. A [product page](/glossary/product-page/) answers someone who wants to buy well and someone who wants to learn badly, and a guide does the opposite.

## How it works

Google makes it the first step of its ranking: **to return relevant results, it first has to establish what you are looking for, the intent behind your query**. It uses language models to connect the few words of the search with the most useful content, corrects spelling and has a synonym system that finds documents even if they do not contain your exact words: "change laptop brightness" finds pages that say "adjust". And it reads context cues: a word like "cook" calls for recipes; "photos," for images; the language decides which language to answer in; a search like "pizza" calls for places nearby; and a topic in the news calls for recent results.[^google-how-search]

### The types of intent, according to Google

The guidelines Google gives the people who evaluate the quality of its results separate two things. The **interpretation** of a query is what it could mean: "apple" can be the brand or the fruit. The **intent** is what the person is trying to get done: someone searching for a movie may want showtimes, reviews or the cast, and the same person can have several intents for a single query.[^google-sqrg] To think about it they propose four types:[^google-sqrg]

| Type            | What the person is after                                                                     | Store example (fictional)                                  |
| --------------- | -------------------------------------------------------------------------------------------- | ---------------------------------------------------------- |
| Know            | To learn or explore a topic. One variant asks for a specific fact that fits in a small space | "how tall should a travel tripod be"; "TV-3 tripod weight" |
| Do              | To get or do something: download, buy, obtain, be entertained or use a website or an app     | "buy TV-3 tripod"                                          |
| Go to a website | To reach a specific website or page                                                          | "Fotoribera online store"                                  |
| Visit in person | To find a business or a type of business to go to                                            | "camera store in Austin"                                   |

Two nuances in the same guidelines prevent misunderstandings. The line between know and do **is often blurry**: the types illustrate general patterns, and many queries do not fit just one. And many queries have several reasonable intents: with "walmart", some want to go to a nearby store, others to shop on its website and others to read news about the company.[^google-sqrg] Those guidelines are not the algorithm: Google makes clear that **no rater's rating changes by itself how a page appears** in Search.[^google-sqrg] They serve to understand what Google considers a good result.

Research uses another, older classification, with three types: informational, navigational and transactional. The paper that gave [GEO (Generative Engine Optimization)](/glossary/geo/) its name, for example, built its benchmark with 80% informational queries and 10% of each of the other two types.[^geo-paper] The two classifications look alike: what changes is the name of the boxes.

### With AI search

Intent does not disappear with [AI search](/glossary/ai-search/); it shows more. Google explains that [AI Mode](/glossary/google-ai-mode/) is especially useful for queries that call for exploring, reasoning or making complex comparisons, and that in it **you can ask nuanced questions that would previously have taken several searches**.[^google-ai-features] A [conversational query](/glossary/conversational-query/) such as "I'm looking for a tripod that fits in a carry-on for landscape photos" carries the intent written into the question itself. What does not change is the advice: Google's systems understand synonyms and the general meaning of what someone is looking for, and **creating a page for every variant of a query**, especially to manipulate rankings, violates its spam policy.[^google-ai-guide]

## On a store

A **fictional example** with a single product, the TV-3 tripod from an invented photography store, and three queries that name it with different needs.

| Query                            | Intent                | The page that answers it               | What that page needs                                                                           |
| -------------------------------- | --------------------- | -------------------------------------- | ---------------------------------------------------------------------------------------------- |
| "TV-3 tripod reviews"            | Know, before deciding | The product page, with reviews in view | The [product reviews](/glossary/product-reviews/), the specifications and an honest comparison |
| "TV-3 or TV-5 tripod for travel" | Know, comparing       | A guide or a comparison                | The differences in weight, folded height and price, with a recommendation by use               |
| "buy TV-3 tripod"                | Do: buy               | The product page                       | Price, availability, delivery time and the buy button                                          |

The table shows what a single page cannot do: serve all three at once without getting in the way of one. **The product page serves the first and the third; the second calls for another page**. What the table cannot say is what Google actually shows for those queries: you only find that out by searching them.


That each query has a single intent and that the way to cover them is one page per query. Google's guidelines say many queries have several reasonable intents and that the types overlap,[^google-sqrg] and its guide for AI search warns that **creating separate content for each search variant can violate its policy against [scaled content abuse](/glossary/scaled-content-abuse/)**.[^google-ai-guide] What fits the intent is the page, not the number of pages.


## How to check it

1. **Search Google for the query you care about** and note what type of pages fill the top results: product pages, categories, guides, maps. It is the best clue to the intent Google reads in it.
2. **Write the intent next to each key page** of your store: know, do, website or visit. If two pages have the same query and the same intent, they compete with each other.
3. **Open the [Search Console](/glossary/google-search-console/) performance report** and look at which queries bring impressions to each product page. A product page that gets "what" or "which" queries is attracting people who have not decided yet.
4. **Check that the page gives what the intent asks for**: to someone who wants to buy, price, availability and delivery; to someone who wants to know, the explanation or the comparison.
5. **Ask the same question in AI Mode, written out in full**, with what you want to get done, and note what type of pages it cites. It accepts nuanced questions that used to take several searches, so it shows which sources it picks when the intent is written into the question.[^google-ai-features]

## Related concept: semantic search

[Semantic search](/glossary/semantic-search/) is **the technique** a search engine uses to interpret the meaning of a query, not just its words. Search intent is **the goal** of the person, what that technique tries to uncover. Google puts them in the same explanation: establishing the intent is the first step, and language models and synonyms are how it does it.[^google-how-search] A store does not control the technique; it controls whether the page answers what the person wanted.

To keep going, the [AI search reading path](/learn/ai-search/) puts the rest of the terms in order.