Search Intent: Four Types, Better Questions, Better Pages

Search intent is the reason behind a search: what someone wants to learn, find, compare, or do. Those four jobs are commonly described as informational, navigational, commercial, and transactional intent. Semrush uses those four categories, and they are a useful starting point for planning a page.

But I think the interesting question comes after the label. If someone arrives on this page, what are they expecting me to help them do?

That question has become more important as I keep building and revising Double Atari. It is tempting to see a keyword opportunity and immediately think, “I should write a blog post about that.” Sometimes I should. Sometimes the better answer is a clearer service page, an honest comparison, or a contact page that does not make someone work so hard.

A keyword tells me what someone typed. Intent helps me decide what to build. Here is how I think about it, which tools I would use, and why the same thinking belongs in an AI search strategy.

Why does search intent matter?

My simplest test is this: does the page finish the job the searcher started? If I search for a restaurant menu and get a long story about the chef's childhood, the writing might be excellent. It is still the wrong answer to my immediate question.

I use the same test for marketing websites. A beautifully written explanation of website audits is not necessarily a good destination for someone who wants to hire an auditor. A sales page asking for a meeting is not necessarily a good answer to “what is an SEO audit?”

Before changing a headline or adding another thousand words, I want to know whether the page has the right job. My recommendation is to make that decision before the content brief, not after a page disappoints.

It also changes how I define success. I would not judge a beginner's guide and a request-a-quote page by the same immediate conversion target. The guide should resolve a question and offer a useful next step. The quote page should help a qualified person act.

What are the four types of search intent?

I translate the four categories into four plain-language requests: teach me, take me there, help me choose, and let me act. These are my shorthand for Semrush's informational, navigational, commercial, and transactional definitions.

The examples are illustrative, not live keyword research or promises about what currently ranks. I would check the actual results before turning any of them into a content assignment.

Informational: teach me something useful

For an informational page, I want the answer near the top, followed by enough detail to make it usable. My schema markup guide for marketers is the kind of content I would assign to this job: explain the concept before asking someone to buy help implementing it.

I would add examples, clarify unfamiliar terms, and offer a relevant follow-up. I would not bury the answer under an introduction that spends six paragraphs announcing how important the topic is.

Navigational: make the destination obvious

For a navigational search, I would make identity and orientation the priority. Someone looking for “Andrew Charon Double Atari” should be able to recognize the right person and business quickly, then get to the About page or the work they came to see.

This is where cleverness can become a tax. I would rather have a clear page title, recognizable business name, and straightforward navigation than a mysterious label that requires an explanation.

Commercial: help someone make a decision

Commercial content is where I want a point of view. My approach is to explain who each option is for, what it costs to maintain in practical terms, and where the compromises show up.

A page like Squarespace vs WordPress for SEO should make the decision easier, not just award a winner. I would rather say “choose the simpler option if your team cannot maintain the more flexible one” than pretend every business needs the same platform.

Transactional: get out of the way of the next step

For a transactional page, I want the offer to be concrete. What is included? Who is it for? What happens after someone contacts me? What information do they need before committing?

I would put that clarity into the SEO consulting service page, not make a reader assemble it from five blog posts. Educational links can still help, but I would keep the primary action easy to find.

Does every keyword fit neatly into one category?

No. A query can carry more than one intent, and Semrush explicitly supports mixed-intent classifications.

Take “SEO audit.” I could imagine someone wanting a definition, a checklist, a tool, or a consultant. I would treat those as hypotheses, then inspect the results and the audience before choosing a page format.

I would also avoid treating “near me” as an automatic purchase signal. For a local business, I would check whether the useful next answer is availability, directions, pricing, credentials, or an explanation of the service. Location is context; it does not remove the need to understand the task.

My rule is to give each page a primary job and a sensible secondary path. A checklist can link to audit services. A service page can answer basic questions. Neither needs to become an encyclopedia that tries to satisfy every possible visitor.

Which tools help determine search intent?

I would combine a keyword tool, a look at actual search results, and first-party performance data. Each gives me a different view; none gets to make the content decision alone.

Google results: inspect the kind of answer being rewarded

I start by looking at the result mix. Are the prominent pages guides, comparison articles, product listings, service pages, or local businesses? I also note the questions, formats, and details those pages emphasize.

I record the date, device, and location rather than treating one search as universal. Google notes that results vary with factors including time, place, device, and recent search history.

Then I ask what I can contribute that is useful and different. Looking at results should help me understand the task, not become an excuse to copy the first five pages.

SEMrush: sort a messy keyword list into a usable starting point

SEMrush labels intent as informational, navigational, commercial, or transactional, and its Keyword Magic Tool can filter ideas by intent. Its documentation explains the labels and the tools that expose them.

I would use those labels to triage a list, then review the actual results for the terms worth pursuing. A spreadsheet label is helpful shorthand, not a substitute for understanding the buyer or checking whether the proposed page makes sense.

Google Search Console: compare the query with the page

Search Console's Performance report provides queries, pages, clicks, impressions, click-through rate, and average position. Google's report documentation describes these dimensions and metrics.

My workflow is to filter to one page, inspect its queries, and ask whether the page answers the questions it is appearing for. If a service page is mostly appearing for beginner definitions, I would investigate the mismatch before simply rewriting its meta description.

I would not call low CTR proof of an intent problem. I would first compare similar queries, positions, devices, countries, and time periods, then inspect the snippet and result layout. Intent is one possible explanation to test, not the only explanation.

GA4: look at what happens after the click

GA4's Landing page report shows where sessions begin and can include sessions, engagement time, and key events; Session source / medium can help distinguish the traffic source. Google documents those capabilities in its Landing page report guide.

I would use this to check whether visitors take the next step the page was designed for. For a commercial comparison, that might be a visit to a relevant service page. For a consultation page, it might be a verified inquiry.

I treat this as supporting evidence, not a mind-reading tool or a reliable way to attach an individual organic keyword to an individual lead. I would also verify the event setup before celebrating the numbers.

AI tools: challenge the brief, then check the answer

I would ask Perplexity or ChatGPT to group queries, suggest alternative interpretations, and identify missing decision criteria. My prompt would include the audience, the current page, and the actual research, not just a bare keyword.

“For each query, suggest the primary intent, a possible secondary intent, the best page format, and the evidence I should check before accepting your recommendation. Flag uncertainty instead of inventing search volume or claiming you inspected results you have not seen.”

That is how I want AI involved: as a thinking partner that makes the brief better. I still want a human to verify the evidence and decide what deserves to exist.

How would I turn intent research into a content plan?

I would start with a small set of pages rather than classify the entire internet. Ten relevant queries and three important pages are enough to test whether the process changes the quality of the decisions.

1. Write down the task. Finish “The person wants to…” without using an SEO acronym.
2. Check the result mix. Record dominant formats, useful details, and signs of mixed intent.
3. Choose the page. Improve an existing URL when it already has the right job; create a new page only when the task is meaningfully different.
4. Write the answer and next step together. Decide what the visitor should understand and what they can do afterward.
5. Choose a measurement. Pair search performance with an action appropriate to the page, then revisit after enough comparable data accumulates.

For a hypothetical audit campaign, I would separate “what is an SEO audit,” “SEO audit checklist,” and “hire an SEO consultant.” I would plan an explainer, a practical checklist, and a service destination, with useful links between them. I would not create a separate page for every tiny wording variation.

This is the missing layer I want in competitor keyword analysis: not just a list of phrases another site ranks for, but a reason for each page I recommend.

Does search intent change for AI search?

I would keep the four categories as a planning framework, but think more carefully about the questions inside a longer request. “Help me choose a CMS for a small marketing team that needs local SEO and cannot support custom development” is much more specific than “best CMS.”

Google says AI Overviews and AI Mode may use query fan-out, issuing related searches across subtopics and data sources while developing an answer. Its AI features guidance explains that process.

My response is to make the decision-making substance easier to find: who an option suits, what its limitations are, which assumptions matter, and what evidence supports the recommendation. I would rather publish one useful comparison than ten pages repeating broad definitions.

This is a content recommendation, not a promise of citations. I would test visibility separately and document what I actually observe, using the kind of process described in my AI visibility tracking guide.

How will I know whether the intent match is improving?

I would keep a simple change log: query group, intended job, page, revision date, and success measure. Then I would compare similar periods and annotate other changes that could explain the result.

For an informational page, I would look for more relevant queries and useful onward journeys. For a service page, I would look for qualified inquiries, not just visits. For a navigational page, I would check whether the intended destination is clear and easy to reach.

The point is not to make every metric go up every month. It is to learn whether the page is attracting the right people and helping them do the thing they came for. That is why I keep coming back to GA4, Search Console, and SEMrush together rather than asking one report to explain everything.

My recommendation: before publishing the next post, write one sentence describing its job. If that sentence is vague, the page probably needs a clearer brief, not more words.

Search intent FAQ

What are the four types of search intent?

The four common SEO categories are informational, navigational, commercial, and transactional. They describe someone wanting to learn, find a specific destination, compare options, or take an action. I use them to choose a page’s primary job, then check the actual search results before deciding on its format.

What is the difference between commercial and transactional intent?

Commercial intent is about evaluating options before a decision; transactional intent is about completing an action. For example, I would use a platform comparison to help someone choose a CMS, and a clearly scoped service or booking page to help someone request implementation support.

Which tools should I use to identify search intent?

My recommended workflow combines SEMrush intent labels, a review of current search results, Search Console query and page data, and GA4 landing-page behavior. Each adds context. I would not let an automated label or a single metric determine the entire content brief.

Can one page serve more than one search intent?

Yes, but I would give it a primary job. An educational checklist can offer a link to professional help without becoming a sales pitch. A service page can explain a concept without turning into a textbook. The next step should support the visitor’s task rather than interrupt it.

Does low click-through rate mean the page has the wrong intent?

Not necessarily. I would compare similar queries, ranking positions, devices, countries, and dates, then inspect the search snippet and competing results. An intent mismatch is one possibility to investigate, not a conclusion I would draw from click-through rate alone.

How should search intent influence AI search content?

I would use intent to identify the specific question or decision a page should resolve, then provide clear explanations, useful comparisons, and honest constraints. That makes the content more useful to readers. It is not a guarantee that an AI system will cite the page, so visibility needs separate observation.