Two queries can share a topic and deserve opposite calls. One wants a free answer, the other wants to pay. Lump them together and the total looks like opportunity. Split them by why and one side disappears.
Search intent is the purpose behind a query (also called user intent or audience intent): what the searcher actually wants to do. Yoast's guide puts it as a plain question: what does this person actually want to do? Guides from Yoast, SE Ranking, and Moz agree on the consequence: matching that purpose decides how the content should be built, before any other optimization matters.
Builders in public threads use that why as the first gate. In one thread about weekly progress, someone laying out a content-first stack describes checking whether there is real search intent, how much demand exists, how competitive the query is, and whether the planned content type matches what the results page rewards, then asking whether the site, at its current domain rating (DR, a third-party score of site strength), can realistically compete. Intent comes first, before internal linking and authority work.
The Four Search Intent Types
The catalog is stable at four types, and the guides agree on the shape.
Informational: learn. The searcher wants an answer, definition, or method. Queries start with what is, how to, why, or best way to. Common examples are what is a gaming computer, how to fix a leaky tap, and what is SEO. The matching format is guides, tutorials, blog posts, and FAQs.
Navigational: reach a known place. The searcher already knows the destination and uses Google as a shortcut instead of typing the full address. Examples include HP website, Trello login, Facebook login, and Yoast SEO. The matching format is the branded homepage, login page, or key product page. Yoast's guide notes the trap here from its own history: its old analytics plugin once ranked for Google Analytics, which sounded like a win, except searchers typing that phrase wanted Google's own site.
Commercial: research before choice. The searcher compares options without being ready to commit. Queries carry best, top, review, compare, or alternatives. Examples include best gaming computers, Asana versus Trello, best coffee maker, and alternatives to Adobe Photoshop. The matching format is comparison guides, reviews, best-of lists, and product roundups that lay out pros and cons.
Transactional: complete the action. The searcher is ready to buy, sign up, or download. Queries carry buy, order, subscribe, download, or the exact product name. Examples include buy HP gaming computer, buy iPhone case, and sign up for Notion. The matching format is product pages, checkout pages, and pricing sheets with a short path from search to action.
Yoast's guide maps Google's own evaluator language, from its Search Quality Evaluator Guidelines, onto the same behavior: Know for learning something, Do for performing an action or purchase, Website for going to a specific site, and Visit-in-person for finding a physical location or business.
Broad queries often blend more than one of the four. In Yoast's description these are mixed intent queries, using running shoes for beginners and best laptops for students under 1000 as examples. In that same description Google answers them with a mixed results page with guides beside reviews beside videos beside product links. Content aimed at a mixed query has to serve more than one need on one page, with clear headings so each group can find its part. The same logic works in reverse when you plan several pages: keywords whose results pages overlap usually share an intent, which is the basis of grouping keywords by SERP overlap.
Search Intent Checker: Read What Google Already Ranks
Every intent checker, whatever it is called, rests on one rule: read the results page Google already shows. If blog posts dominate, build an article; if product pages dominate, build a landing page. Yoast's guide spells out the signals: guides and how-to articles mean informational intent, reviews and comparisons mean commercial investigation, and product pages mean transactional intent.
Run it as three looks at one SERP, the page of results Google shows for the query:
- What format holds the top. Note what actually ranks: tutorials, comparison tables, product pages, videos. Deliver the format the top already rewards. Yoast's guide uses a food query to make the point: someone typing tomato sauce usually wants recipes rather than a history lesson, and Google ranks accordingly.
- Whether the page mixes. If guides sit beside product links, treat the keyword as mixed and plan content that answers the informational question, includes the comparison, and points to the action, rather than forcing one shape.
- Whether the answer comes first. The pages that hold the top usually answer the core question in the first sentence or section and add detail after. Do the same.
Yoast's guide adds the audience question that keeps the read honest: what does someone hope to achieve when they type this keyword. Two people can type the same words and expect different things, which is why the same guide recommends checking analytics search terms, customer questions, and a short on-site poll alongside the results-page read. Labels help triage. The Yoast SEO and Semrush integration described in Yoast's guide tags related keywords I for informational, C for commercial, T for transactional, and N for navigational, and keyword tools such as Semrush and DataForSEO offer intent labels per keyword. Use a label as a starting sort and let the live page have the last word. Pulling that page takes one call to any live-results API; qequals, for one, returns the current Google results for a query, which is all this check needs.
When Intent Says Do Not Enter
Volume totals hide intent splits. Splitting them changes the call.
In a long public research teardown that calibrated guesses against real DataForSEO volumes, the author separated pure-fortune demand from AI-times-fortune demand in Thai search. The pure-fortune family ran about 1.95 million monthly searches across a 6,420-word family; the AI-times-fortune family ran 8,710 across 10 words, a ratio around 224 to 1. That separation changed the whole research direction: ChatGPT fortune-telling stayed an observation, while fortune-telling itself was the pool.
The same teardown then split free intent from paid intent. Free fortune-telling drew about 8,100 searches; price and paid-booking phrasings drew 30, 20, zero, and zero. At roughly 160 to 1, the numbers closed the subscription path in one comparison.
That filter sequence is the pattern to copy: real intent, then demand, then competition, then whether your current site can realistically compete. Another public workflow states the validation step in one line: after mapping competitor sitemaps to assumed keywords, verify unique search intent by searching the keyword itself. A related workflow adds volume through a keyword API and sorts the content roadmap by intent first and volume second, so a large learning-only pool never outranks a smaller buying one by accident.
Ask the kill questions in order. Does the keyword carry buying intent or only learning intent. Does the free-versus-paid split leave anyone willing to pay. Does the results page reward a format you can credibly build. Can your site, at its current strength, displace what already holds the top. A yes on volume with a no on any of those is a no on the niche.
Date each read. Intent is stable for most queries, but the page that expresses it is not, and a keyword that rewarded guides in spring can reward product pages by the holidays.
