What is search intent?

Search intent is the reason behind a search: what someone wants to find or accomplish when they type a query into Google. RankDots recognises five intents - Informational, Navigational, Commercial, Transactional and Local - and a keyword can carry more than one. The intent tells you what kind of page has a realistic chance to rank.

Where you’ll see it

• Keywords and Pages screens: the Search intent column, with a letter badge for each intent a keyword or page carries. A keyword shows one badge, or two when its top two intents are close. You can filter by it.
• Each topic cluster card and a page’s details: Intent, a bar showing the share of each intent. The shares always add up to 100%; point at the bar to see the percentages.
• Content settings, when you create content: Search intent. RankDots pre-selects the intents that match the topic, and you can choose up to three.

The five intents

How to read it

Write for the intent, not just the keyword. Google ranks pages that give searchers what they came for. A detailed guide rarely ranks for a keyword where people want to buy now, and a product page rarely ranks where they want to learn. Matching the page type to the intent matters as much as covering the words.

Several badges mean mixed needs. Decide which need comes first and build the page around it. The Intent bar shows which intent carries the most weight.

Intent hints at value. Commercial and Transactional searches sit closer to a sale; Informational searches build reach and trust earlier on. Neither is better in general - it depends on what you need now.

Navigational searches for other brands are rarely worth chasing. Someone looking for a specific company wants that company’s site.

How we calculate it

RankDots works out intent as probabilities. For each keyword, it estimates how likely the search is to be Informational, Navigational, Commercial, Transactional or Local, using two sources:

• Keyword modifiers. Words in the query that signal intent. A query containing “how to” is more likely Informational; one containing a site’s name, such as “YouTube”, is more likely Navigational. When a query has no recognised modifier, this source uses a default: 60% Informational and 10% for each of the other four.
• An intent model. A language model based on BERT, trained on search queries already labelled by intent, reads the whole query.

Each source gives five percentages. RankDots averages them, intent by intent. The keyword is labelled with the intent that has the biggest share - or with the top two, when their shares are close to each other.

A page’s and a topic cluster’s intent come from their keywords: the keywords’ intent percentages, weighted by each keyword’s search volume. High-volume keywords therefore shape the Intent bar most.

Search intent is worked out when your project is created and is not refreshed afterwards.

What it doesn’t tell you

• The certain intent of a query. Intent is an estimate. For a query with no telling words, half of the estimate starts from a default that leans Informational, so read its label with care.
• How intent has shifted since you created the project. The labels show the state at creation.
• How many people search. Search volume does that.
• Whether you can rank. Weak spots and keyword difficulty answer that.
• The exact format Google rewards right now. Look at the pages in the Google Top 20 column, and at the reference pages in Content settings.
• Anything about AI Visibility prompts. Prompts carry their own Intent, with different values, and it is a separate label.

What next

• Step 1 — Content settings [link is coming]
• How to choose which topic to work on first [link is coming]