What intent is
Behind "best running shoes" is a person about to compare and buy; behind "how to lace running shoes" is a person with shoes and a problem; behind "nike" is a person going somewhere specific. The words overlap; the needs do not. Search engines have spent decades learning which results people are satisfied by for each query, and ranking is in large part a prediction of satisfaction. A page's job is to be what satisfies.
Google's documentation on helpful content frames the same idea from the writer's side: content made to serve people's needs, that leaves them satisfied, that matches what they came for. Intent is the operational version of that principle — the specific need behind the specific query.
The four types and their limits
The common classification has four kinds of intent:
| Intent | The person wants to… | Typical queries | Typical satisfying page |
|---|---|---|---|
| Informational | learn or understand | "how to", "what is", "why does", questions | Guides, explainers, tutorials, definitions |
| Navigational | reach a specific site or page | brand names, product names, "login" | The destination itself |
| Commercial investigation | compare before buying | "best", "vs", "review", "top", "alternatives" | Comparisons, reviews, round-ups, buying guides |
| Transactional | act now: buy, sign up, download | "buy", "price", "near me", "discount", product plus modifier | Product, service, pricing and landing pages |
The classification is useful for planning — the intent classification guide uses it to sort keyword lists at scale — and it is too coarse for optimising a single page. "Best running shoes for flat feet" is commercial investigation, but the results page tells you whether people want a list of ten, a detailed review of three, a page from a retailer, or expert guidance from a podiatrist. The type says compare; the results page says how. For stores, the guide to buying guides for online stores covers the commercial investigation format in depth, including how guides and category pages divide the queries between them.
Reading intent from the results page
The method is to search the target query (in the right country and language, in a clean browser profile) and read the page as evidence.
Page types that rank. Count what appears in the top results: guides, product pages, category pages, comparisons, videos, forums, tools. The dominant type is the format the engine rewards. A mix signals mixed intent (below).
Features on the page. Shopping units and product ads signal transactional intent; a featured snippetA highlighted extract from one page shown as a direct answer in search results. and people-also-ask boxes signal informational; a local packThe map and short list of business listings shown for searches with local intent. signals "near me" intent even without the words; video carousels signal that people want to watch; image blocks signal visual research.
Angle and depth of the ranking pages. Open the top few. How long are they, what do they cover, what do they assume the reader knows, what do they lead with, what do they link to? Beginner-level explainers and expert reference pages rarely rank together; the results show which is wanted.
Titles and snippets. The words the ranking pages use in titles are a record of what the engine chose to show for this query; patterns ("guide", "checklist", "prices", "2026") reveal the framing searchers respond to.
Freshness. Dates on the results indicate whether the query wants current information (news, prices, annual lists) or evergreen material.
Ten minutes with the results page answers most intent questions better than any tool.
Where the results carry video and the business runs a YouTube channel as well as a website, the tutorial on planning YouTube videos and web pages for the same searches turns this reading into a decision about whether a page, a video or a linked pair should own the query.
Matching the page
Once the intent is read, the page is matched on four dimensions:
- Format. If the results are comparison tables, the page needs a comparison table, not a narrative. If they are step-by-step tutorials, the page needs steps. Format is the first filter and the one most often wrong.
- Depth and level. Match the length and expertise of what ranks. A short page competing with long references will lose; a long page competing with short answers will not be read.
- Angle and lead. What the person wants first: the answer (put it first), the comparison (lead with it), the price (show it). Burying what the query asked for under preamble fails the intent even when the content is present.
- Completeness. The sub-questions the results answer (visible in people-also-ask and in the ranking pages' headings) are what the query implicitly contains; the page covers them. The headings guide turns these into a structure.
The content brief tutorial captures all four as the brief for a writer, which is where intent matching becomes repeatable rather than a reviewer's instinct. Queries that name two products, or ask for the alternatives to one, have a page type of their own, covered in the guide to comparison and alternatives pages.
Two pages for one query
A kitchenware retailer wants to rank for "cast iron pan seasoning". The results page shows step-by-step guides from cookware brands and publishers, a video carousel, and people-also-ask questions about oils, oven temperature and rust. No product pages rank. The retailer's existing page for the query is a product category with a paragraph of tips at the bottom — the wrong format, buried angle, incomplete coverage. The page that ranks is a tutorial: numbered steps first, the oil and temperature questions answered under their own headings, an embedded video, and a link to the product category at the end for people who want to buy. The category page keeps its own query ("cast iron pans"), where the results are, in fact, category pages.
Mixed and ambiguous intent
Some queries have no single intent — "python" (the language or the snake), "apple watch bands" (buy or compare), "marketing plan" (learn, template or tool). The results page for these shows a mix, often with the dominant interpretation first. Three responses:
- Target the dominant interpretation and accept the share of searchers it covers.
- Serve the secondary intent within the page where it is a natural extension (a comparison that ends with where to buy), without turning the page into two pages.
- Create separate pages for genuinely separate intents and link them, rather than forcing one page to do both. The cannibalisation guide covers the opposite failure, where separate pages end up competing for one intent. When a live page already tries to serve both, the tutorial on splitting a page by search intent divides it into two without losing the links the original earned.
Fractured intent — where the results page is a mix because the engine itself is uncertain — tends to resolve over time toward one format; watching the page over months shows which way.
Intent drift
Intent is not fixed. Queries that once wanted definitions come to want comparisons as a category matures; queries that wanted comparisons come to want products as a brand becomes known; seasonal queries change intent through the year; a news event temporarily turns an informational query into a navigational one. The engine follows searcher behaviour, and a page that stops matching the results page will lose position without having changed. The defence is monitoring: for the pages that matter, re-read the results page on a schedule and when a ranking moves without explanation, and update the page's format and angle when the intent has moved. The SEO KPIs guide covers tracking position and click-through (CTRShare of impressions that result in a click.) by query, which is where drift shows first.
Intent and the customer journey
Search intent is one view of a larger idea: people reveal where they are in a decision through what they do. The intent signals guide in the fundamentals cluster covers the same reading across channels — queries, pages viewed, actions — and maps the four search-intent types onto journey stages. The See-Think-Do-Care framing is the same insight applied to channel planning. For SEO the practical point is that a site needs pages for each stage's intents, not only for the ones that convert, which is what the topic clusters guide plans.
Recording the reading
An intent reading is only useful if it survives the person who made it. Record, per page: the head query; the date; the dominant intent and format; the features present; the angle the top results take; what they miss; and the decision made for the page. Keep it with the content brief, and re-read the results page when the page is refreshed, when its click-through falls without a position change, or at the quarterly research review. Half of intent problems are pages built correctly for a results page that no longer exists.
When a page loses positions because the results now favour a different kind of answer, the sequence on diagnosing a sudden rankings drop helps rule out technical and reporting causes first.
An intent mismatch is the first cause checked in new content that is indexed but not ranking, followed by cannibalisation, thin content, internal links and topic depth.
Common mistakes
- Deciding intent from the words. "Best" does not always mean a list; the results page knows.
- One page for every intent. A product page with a buried tutorial, or a tutorial that is secretly a sales page.
- Matching the words, not the format. Keyword presence without the table, steps or answer the query wants.
- Writing for the wrong level. Beginner content for an expert query, or the reverse.
- Ignoring the results page's features. A video carousel and a local pack are intent evidence.
- Never re-reading the results page. Intent drifts; pages that do not follow it fade.
The intent classification guide covers sorting intent at the scale of a keyword list; the author's on-page SEO expertise page describes how intent checks fit the working process.
