Search Intent: Matching Content to What People Actually Want

Two people can type the same four words into Google and want completely different things. One wants a definition. One wants to buy something in the next ten minutes. Search intent explained simply is this: the job behind the query, not the words used to express it. Get the job right and a modest page performs. Get it wrong and a well-written page fails anyway, because it answers a question nobody asked.

That's the pattern worth sitting with before anything else. This isn't a wording problem you fix with better synonyms. It's a targeting problem, and it happens upstream of the sentence level entirely.

The real job behind a query

Search intent is the outcome a person wants when they search, not the string of characters they used to search for it. Google's own guidance on helpful content asks whether a page genuinely satisfies the person who lands on it, whether it offers something they'd feel was worth their time, rather than whether it hits a keyword (Search Central). That's the whole test. Not "does this page mention the topic" but "does this page finish the reader's task."

Diagnosing intent wrong is probably the single most common reason strong content underperforms. The topic is relevant, the research is solid, the keyword sits right there in the H1, and the page still goes nowhere. It's not a writing problem. It's answering the wrong kind of question. A definition where the reader wanted a shortlist. A sales page where the reader wanted to learn something first. The prose can be excellent and it still won't matter, because relevance gets judged on task completion, not on craft.

There's a diagnostic signal for this in your own data, if you know where to look. Google's Search Console documentation notes that a low click-through rate on a page that's already ranking can mean people saw it in the results and decided it probably wouldn't answer their question (Search Console Help). That's intent mismatch showing up as a number, not a theory. If a page ranks on page one for the right keyword and still gets ignored, the snippet is telling searchers something the content later confirms: wrong shape for the job.

This is also why intent has to be the first decision in a content plan, not a detail settled after the draft is written. Get the job wrong at the brief stage and everything built on top of it, the structure, the examples, the length, is aimed at the wrong target.

Four intent types

Most queries lean on one of four jobs: learning, navigating, comparing, or acting. Knowing which one you're dealing with tells you almost everything about how the page should be built, before you write a word of it. Each one also maps to a different stage of the reader's journey, from first noticing a problem exists to actually paying for a solution.

Intent	Reader's stage	Example query	What the page needs
Informational	Just realised there's something to learn	"what is search intent"	A clear answer, near the top, no padding
Navigational	Already knows the destination	"Asana pricing"	To be the destination, or get out of the way
Commercial	Comparing options before deciding	"Notion vs Asana"	Real criteria, honest tradeoffs, a verdict
Transactional	Ready to act right now	"sign up free trial"	Minimal friction, fast route to the action

Informational queries want knowledge. "What is search intent," "how to structure a blog post," "why do backlinks matter." The reader wants a clear answer, ideally near the top, with enough depth to actually be useful and none of the padding that exists just to hit a length. Ahrefs' review of AI Overviews found that informational queries make up the overwhelming majority of the sample they studied in the US, far outweighing navigational or transactional intent (Ahrefs). Most of what gets searched, in other words, is someone trying to understand something, not buy it. This sits early in the journey, before the reader has even decided a solution is worth pursuing.

Navigational queries are looking for a specific destination. A brand name, a login page, "Asana pricing." You generally only win these if you're the thing being searched for. Trying to intercept someone's navigational query with a generic explainer rarely works, because the reader already knows exactly where they're headed and your page isn't it. This one sits wherever the reader already is in their journey, since it's really about locating something they've already decided to use.

Commercial queries sit right before a decision gets made. "Best project management software," "Notion vs Asana," "top CRM tools." The reader is comparing, and wants criteria, honest tradeoffs, and something close to a verdict. Semrush's keyword research materials treat this as one of the four standard categories worth planning around deliberately, distinct from pure informational search (Semrush). A neutral encyclopedia entry doesn't satisfy this reader. Neither does a page that's obviously just trying to sell them one option. This is the late-consideration stage, right before the reader narrows their options down to one.

Transactional queries mean the reader is ready to act. "Buy," "sign up," "pricing," "free trial." They want the fastest possible route to doing the thing, with as little reading as possible standing in the way. This is the decision stage itself, and anything that slows the reader down here works against the page rather than for it.

Ahrefs' work on HubSpot is a useful reminder that these categories aren't always clean. A single page can serve dual intent when the format and on-page elements handle both jobs at once, learning and signing up, in one place, because Google evaluates the whole page rather than a single dominant purpose (Ahrefs). Say a landing page opens with a clear explanation of what a tool does, then moves into a signup form. That's not intent confusion. That's a page built for a reader who's already most of the way decided.

How to read the SERP

The query is a hypothesis about intent. What's actually ranking, and what AI answer engines are surfacing, is the evidence. When the two disagree, trust the evidence.

Start with the query itself for a first guess. Signal words point somewhere before you've checked anything else. "How to" and "what is" suggest informational. "Vs" and "best" suggest commercial. "Buy" and "near me" suggest transactional. That's a fine starting hypothesis. It is not a conclusion, because plenty of queries carry more than one plausible reading and phrasing alone won't settle which one wins.

The stronger signal is what the results already show you. Google has noted that many different phrasings, misspellings, and even different languages can map onto essentially the same underlying intent, which is part of why matching the literal words matters less than matching the job those words are pointing at (Search Console Insights docs). The engine has already resolved that ambiguity by watching what real searchers chose, again and again, across every variant of the phrasing. That's a stronger data point than your read of one query string.

A rough methodology, if you want to make this repeatable: pull up the query in an incognito window, note the top eight to ten organic results plus any AI Overview or featured snippet, and log format, depth, and angle for each. Do this before you write, not after, so the brief reflects what's live rather than what you assumed six months ago.

Look at four things in the live results.

• Format. What dominates: listicles, step-by-step guides, product pages, comparison tables, video?
• Depth. Is the typical answer two paragraphs or a genuinely comprehensive guide?
• Angle. Are the top pages educational and neutral, or comparative and opinionated?
• AI treatment. Does the AI answer resolve the query directly in a couple of sentences, or cite several sources because the topic needs more nuance than one page can give?

That last point matters more now than it used to, and it's a big part of what generative engine optimization is actually about: understanding whether you're competing to be the quick answer or to be one of several sources cited alongside others.

Here's how that plays out on a real query. Take "how to structure a blog post." Run the check above and the SERP is dominated by numbered, skimmable guides with headers every 150 to 200 words, several with a jump-to-section table near the top, and the AI Overview summarises a short list of steps rather than citing one dominant source. That pattern tells you the winning format isn't a single flowing essay. It's a scannable how-to with clear subheads and a structure the reader can skip through, because that's what the whole first page has already agreed on.

Ahrefs has also documented that dominant intent for a keyword can shift over time, causing a page to lose rankings with zero changes made to the page itself, because the meaning or dominant purpose behind the query moved and the content didn't move with it (Ahrefs). Intent mapping isn't a one-time exercise you finish and file away. Queries that meant one thing two years ago can mean something else now, and the only way to catch that is to keep checking the live results, not just trust the mapping you did when you first built the page.

Match format and depth

Once the intent is clear, format and depth should follow from it almost automatically. The job is to hand the reader the shape of answer their intent is calling for, not the shape you'd prefer to write.

Informational intent wants a clear explainer or how-to, answer near the top, headings that let someone scan to the part they need, enough depth to be genuinely useful without padding for the sake of a word count. Commercial intent wants a comparison with real criteria and an honest verdict, not a page so hedged it never actually says anything. Transactional intent wants friction stripped out entirely, a fast path to the action, no essay standing between the reader and the button. Navigational intent wants the actual destination, delivered fast, without burying what they came for under content they never asked for.

Depth isn't a separate knob from format, it's part of the same match. Ahrefs' AI Overviews study found that the pages cited most often for complex informational queries tend to run longer and cover more subtopics than the pages cited for simple ones, which lines up with what you'd expect: a thin page loses to whoever actually went deep when the query demands it (Ahrefs). Over-write a simple query and you bury the two-line answer someone wanted inside six paragraphs of throat-clearing. Under-write a genuinely complex one and the reader bounces straight back to the results, unsatisfied, looking for whoever actually went deep.

Say someone searches "what is a sitemap." That's a two-paragraph answer, not a 2,000-word guide. Say someone searches "technical SEO audit checklist." That's the opposite problem, and a thin page will lose to whoever wrote the comprehensive one. Calibrate to the question in front of you, not to a target length decided in advance.

Where pages miss

The most damaging mismatch isn't weak prose. It's the wrong page type for what the reader came to do, and that failure happens regardless of how well the sentences are put together.

Selling to a learner is the most common version. A product page or a hard pitch shows up for a query that was clearly informational, and the reader who wanted to understand something leaves without reading past the first paragraph. Take "what is a CRM." A vendor page built for that query, once it opened with a features list and a demo booking form, was pulling low single-digit click-through despite ranking on page one, because searchers could tell within seconds it wasn't going to teach them anything. Rebuilt as a plain explainer with the pitch moved to a single link near the bottom, the same query started holding attention past the first screen. Same topic, same keyword, different job.

Lecturing a buyer runs the opposite direction. A neutral, textbook-style page shows up for a query that was clearly commercial, where the reader wanted a verdict and got an encyclopedia entry instead, so they go find the shortlist somewhere else. The brief difference matters here: a page built for "what is project management software" only needs a clear definition and maybe a short history. A page built for "best project management software" needs named products, a comparison table, and a stated pick, because the reader isn't trying to understand the category anymore, they're trying to choose from it. Writing the second brief like the first is the whole failure in miniature.

A thin answer sitting where every competing result is a comprehensive guide will lose, and a bloated 2,000-word essay sitting where the reader wanted one clear sentence will lose too, just from the other direction. Wrong format entirely is its own failure, a wall of prose where every other result on the page offers a comparison table or a numbered process, forcing the reader to do work that competitors handed over for free.

The signal these mismatches leave behind is usually consistent: decent impressions, weak engagement. The page shows up in search, people click, and they leave fast, because they recognized within a few seconds that this page wasn't built for their job. That's rarely a writing fix. It's a format fix, and it means going back to the intent decision, not polishing sentences that were never going to convince anyone.

Intent should shape more than a single page, too. It's the logic that should sit behind internal linking and cluster structure, because grouping content by topic without accounting for intent tends to link informational pages to transactional ones in ways that confuse both the reader and the crawler. A cluster built around one topic but three different jobs needs three different kinds of pages, linked deliberately, not just tagged under the same category.