- Classify each target query as informational, navigational, commercial, or transactional before drafting, then match the format that dominates the first page of results.
- Confirm intent from the live results by labeling the top ten by page type in a fresh session in your target market, and save a dated screenshot of what you saw.
- If a page answers a neighboring question, rewrite it to match the query instead of adding keywords, because a mismatched page rarely ranks and tends to attract the wrong visitors.
A page can be accurate, well written, and technically clean, and still never rank for the query it was built for. The cause is rarely the writing. The page answers a different question than the one the searcher typed. Someone searches “project management software for agencies” and lands on a 2,000-word explainer about what project management is. The explainer may be excellent. It is still the wrong page, and the results page was assembled around that gap before you wrote a word.
That gap has a name: search intent. It is the job a searcher is trying to get done when they type a query, and it decides which kind of page belongs in the results. Match it and your format, depth, and calls to action mostly fall into place. Miss it and no amount of keyword work will rescue the page.
This article covers the four intent types, how to read the live results to confirm which one applies, and how a single mismatched page fails. Most of that failure is quiet, which is why teams tend to blame everything except the page type.
What search intent is, and why keywords alone don’t tell you
Search intent is the reason behind a query. The words are a compressed signal: a few terms standing in for a goal the searcher never states. Keyword tools report volume and difficulty for the string, but nothing about the goal behind it. “Trail running shoes” might come from someone ready to buy a pair, someone comparing brands, or someone who doesn’t yet know what a heel drop is.
Google’s job is to guess that goal and return what best serves it. You can see the guess when you search. The results are not a random sample of pages about trail running shoes. They are a bet on what the typical person typing those words wants, and changing the words changes the bet. “Best trail running shoes for flat feet” pulls roundups. “Trail running shoe deals” pulls retailers. “How to choose trail running shoes” pulls guides.
The framework most SEO practitioners still use starts with Andrei Broder’s 2002 paper “A Taxonomy of Web Search,” which sorted queries into navigational, informational, and transactional. Commercial investigation, the research step before a purchase, was added later as a fourth category, and it earns its place because it behaves differently from both pure learning and pure action. Four types is the working standard. It is useful because it is blunt. You don’t need a perfect taxonomy. You need to know which format to build.
The four types of search intent, and the format each one wants
Informational
The searcher wants to understand something or get an answer. “What is a canonical tag,” “how long to rest a steak,” “why does my sourdough collapse.” Informational queries are the largest group on most sites and the one teams write for most often, which also makes them the easiest to write badly.
The format that wins answers fast, then goes deeper. Put a direct answer in the first few lines, explain the mechanism underneath, and show one worked example. A searcher asking what a canonical tag is wants the definition first, then the reason it exists, then a case where it matters. The common failure runs the other way: a brand story or product pitch sits on top of the explanation. The searcher leaves, and the page earns neither the answer nor the trust.
Navigational
The searcher wants a specific site or page. “GitHub login,” “HMRC VAT registration,” “Stripe dashboard.” The brand or destination is in the query, and the person already knows where they are going. The official destination almost always ranks first, and everything else on the results page is noise to them.
That makes navigational intent a strategic filter more than a content format. A third-party blog rarely displaces the official site for someone else’s brand, and effort spent chasing those queries is effort spent on a result the searcher doesn’t want from you. Where navigational intent matters is your own brand. Your login, pricing, and documentation pages should be the obvious match for branded queries, and you should check that they are.
Commercial investigation
The searcher is comparing options before a purchase. “Best CRM for small agencies,” “Notion vs Coda,” “[product] review.” They have a problem and a rough budget, and they want help narrowing the field. Commercial queries reward honesty more than any other type. Readers are looking for criteria and a verdict, and they can tell when a page ranks its own product first and calls it the best.
The format that fits is a comparison built on the criteria a buyer actually uses: price, setup time, team size, integrations, limits. A table earns its place here. So does a plain statement of who each option is for and who should look elsewhere. An explainer that never names a real option will fail this query no matter how well it is written.
Transactional
The searcher wants to act now. “Buy trail running shoes,” “plumber near me,” “download Figma,” “cheap flights to Lisbon.” The decision is made or nearly made, and the only thing between intent and action is a page that gets out of the way. Transactional results favor product and category pages, visible prices, stock and delivery details, and one obvious next step. Long prose is friction. A transactional page with a 1,500-word essay above the price is working against the person who came to buy.
The line between commercial and transactional is soft. Someone reading reviews of a laptop may be one click from checkout, and the results will show a mix of review and shopping listings. Treat the two as ends of a range, and let the results show you where a query sits on it.
Here are all four side by side. Read the last column first, because that is where most lost traffic comes from. Teams usually classify queries correctly in their heads and then publish the wrong format.
| Intent | What the searcher wants | Typical query shape | Format that fits | Typical mismatch |
|---|---|---|---|---|
| Informational | An answer or an explanation | “what is”, “how does”, “why” | Direct answer, then mechanism and examples | Product pitch with no explanation |
| Navigational | A specific site or page | Brand names, “login”, “[brand] pricing” | The destination itself | Third-party blog about the brand |
| Commercial | Evaluate options before buying | “best”, “vs”, “review”, “for [use case]” | Comparison with criteria and verdicts | Explainer that names no options |
| Transactional | Take an action now | “buy”, “price”, “near me”, “download” | Product or category page, price, clear next step | Long essay before the price |
Mixed intent is real, and the results page shows it
Plenty of queries carry two intents at once. “iPhone 16” could mean “what does it cost,” “is it worth upgrading,” or “what are the specs.” “Standing desk” spans guides, category pages, and retailers. Google’s ranking systems handle this by mixing page types in the results, so you will often see a category page beside a review beside a reference article.
For a mixed query, match the dominant format, then decide whether the minority intent deserves its own page. Don’t try to serve both on one URL by stacking a guide above a product grid. The page ends up too long for buyers and too thin for learners. Two focused pages that each match one intent usually beat one compromise, and they give you two chances to rank instead of one.
Intent also moves over time. A category new enough to need explaining will eventually attract comparison queries as buyers arrive, and the results page reflects that shift as it happens. A query you classified as informational last year may now be commercial. That is why the SERP check belongs in every content refresh, not only at launch.
How to read the SERP to confirm search intent
Your keyword tool gives you a guess. The results page is evidence. Google has spent years tuning its results to what searchers actually accept, so the top ten is the closest thing you can see to an answer key. Here is the process we run before anyone writes a word. Start from a list of target queries, such as the one that keyword research in the AI era produces, and run these steps on each one.
- Search the exact query in a fresh session. Use a private window in the country and language you target. Personalization and location change results, and a page of results from your logged-in laptop can mislead you about what your audience sees.
- Classify the top ten by page type, not by topic. Label each result as a guide, product page, category page, comparison, review, forum thread, video, or official homepage. Two pages on the same subject can have opposite intents, so label the format, not the subject. When eight of ten are comparisons, you have your answer.
- Read the SERP features as hints. Shopping units signal transactional intent. A “People also ask” block shows the sub-questions the searcher is working through. Video results suggest they want to see something done. An AI Overview gives Google’s summary and the sources it drew from, which is worth reading closely because it shows what a complete answer looks like to the system. Local results mean the person wants a place nearby.
- Compare structure, not just type. Look at the top three. Do they share a length range, a heading pattern, a table, a set of named products? Shared structure is what the expected format looks like. If the top three are all 2,500-word roundups with comparison tables, a 900-word essay will struggle, however good it is.
- Write the decision down in one sentence. Something like: “Commercial intent. Buyers want a shortlist of five tools compared on price and team size, with a verdict.” If the format you want to write doesn’t match what ranks, you have three honest options: write the format that ranks, pick a different query, or accept that you’re targeting a long shot and budget for it. Skipping the decision is the expensive option.
Tip: Save a dated screenshot of each results page you check. Results shift after core updates and product launches, and a dated record is the only way to tell a real change in intent from a one-off fluctuation.
How one mismatched page fails
Mismatch failures are slow and quiet. Nothing errors. The page is indexed, it may even rank, and it produces visits that go nowhere. The clearest way to see the pattern is a worked example. This is a composite scenario rather than a client case, but the sequence is common.
A B2B software team wants to rank for “project management software for agencies.” Their writer produces a 2,200-word article titled “What is project management? A complete guide.” It is well researched. It covers the history of the discipline, the five process groups, Gantt charts, and the critical path. It is indexed within days and linked from the blog.
Then the numbers arrive, and they look like this:
- For the target query, the page sits well outside the top ten. The results are comparisons with named tools, pricing, and criteria. This page names no tools and gives no criteria, so it never enters the consideration set.
- For a nearby query, “what is project management,” the page ranks and draws steady traffic. That traffic is real. But the visitors are students and people new to the field, and almost none of them are agency owners who would buy. The dashboard climbs, which feels like progress.
- The team sees a flat conversion rate and concludes that SEO doesn’t work in their category. They commission three more explainers. Each one adds to the same mismatch, so the site now has four pages competing for learning queries and none competing for the buying queries that pay.
That last step is the expensive one. The first mismatch was a single decision. Repeating it compounds the cost, and the post-mortem blames the channel instead of the page type. Every part of this failure was visible before the first draft existed. A short SERP check would have shown the comparisons.
Why does one page fail this way? Authority gets a page into the candidate set for a query, and intent decides how it is ordered inside that set. A strong, well-linked page that answers a neighboring question has low relevance to the query it targets, and more links rarely fix that. The ranking systems are trying to surface the pages that best satisfy the query as written. A page that answers a different question is competing without the thing that counts.
Warning: Ranking is not proof of a match. A mismatched page can hold a middling position for months while sending visitors who can't do anything with it. Judge the fit by what the visitor does next, not by where the page sits.
The fix for this team was not more content. It was a rewrite of the target page as a comparison: a criteria table covering per-seat price, time tracking, client portals, and resource planning; five named tools, each with a plain “best for” line; and an honest note on where their own product fits and where it doesn’t. The explainer moved to its own URL and got linked from the comparison as an aside for readers who are new to the field. Then the team checked the results again before deciding what came next.
Matching format to intent when you write
Once you have confirmed the intent, the writing decisions get simpler. Work through them in order, because each one depends on the one before.
- Put the matching answer in the first screen. Informational pages give a direct answer in the first two or three sentences. Commercial pages give the shortlist or verdict before the detail. Transactional pages put the product, the price, and the action above the fold. Navigational pages put the destination link where the eye lands first.
- Set depth to what the results reward. If the top results average 2,000 words of comparison, a 600-word post will struggle to match them. If they are 400-word definitions, a 3,000-word essay is effort the reader won’t use. Length is a symptom of intent, not a goal in its own right.
- Use the structure the results use. Tables for comparisons, numbered steps for procedures, specs for products. Headings should mirror how people phrase the question. A heading that answers a question someone actually asked is easy to scan and easy to quote.
- Give the next step the intent calls for. Informational readers want the next explainer or a tool. Commercial readers want the detail behind the verdict or a trial. Transactional readers want checkout. A single call to action that ignores intent is a second mismatch, sitting on the same page as the first.
- Brief before you draft. Record the intent, the format, the required sections, and the SERP evidence in a brief the writer works from. That document is where a mismatch gets caught before it costs anything. The SEO content brief guide walks through how to structure one.
What this means for your content plan
Intent is the cheapest SEO work you will do and the most often skipped, because it feels like planning rather than progress. Here is what I would insist on. Classify every query before anyone drafts. Run the results check and write down the decision. Refuse to write a page you cannot match honestly.
Not every query is worth chasing. If a query’s intent is navigational to a competitor, or the transactional results are dominated by marketplaces with budgets you can’t match, the honest move is to spend your effort on a neighboring query where you can match the format. A site that matches intent on thirty queries will usually outperform one that loosely covers three hundred. Breadth without a format match is the programmatic trap in a different costume.
The same logic is showing up in AI answers. Systems that generate answers tend to draw on passages that address the question directly, so a page that answers a neighboring question is less likely to be the source they quote. Matching intent serves both the ranked result and the cited answer.
That is the discipline AutoSEO Stack automates. It researches the queries your buyers actually use, writes each article in a human voice against a specific query, and publishes on a cadence with the per-page quality bar built in. If you want to see how that research looks on your own site, start with the Google research view or take a look at the features overview.