What Is Search Intent? The Four Types, and Our Data Showing Intent Does Not Predict Clicks

Search intent is the goal behind a query. Someone typing "running shoes" wants to browse, someone typing "best running shoes for flat feet" wants to be advised, and someone typing "nike pegasus 41 size 10" wants to buy. Same topic, three different pages required.

That much is settled and every guide on this keyword says it. What almost none of them do is check whether matching intent actually produces traffic. We pulled 12,145 queries from our own Search Console and measured it, and the answer is more complicated than the framework suggests.

The four types, defined properly

Informational. The searcher wants to know something. "What is search intent", "how to change a tyre", "population of Portugal". The largest category by query count on most sites. Satisfied by guides, explainers, definitions and tutorials.

Navigational. The searcher wants a specific site and is using the search box instead of the address bar. "Gmail login", "distribb pricing", a brand name on its own. You cannot win these for someone else's brand, and you should never try.

Commercial. The searcher intends to buy but is still deciding. "Best CRM for small business", "ahrefs vs semrush", "surfer seo review". Pre-purchase research. Satisfied by comparisons, listicles and reviews.

Transactional. The searcher is ready to act. "Buy standing desk", "semrush pricing", "book dentist appointment". Satisfied by product pages, pricing pages and booking flows.

There is a fifth case worth naming because it causes most misclassification: mixed intent, where one query genuinely serves two purposes and the results page shows both. "Content calendar template" returns free downloadable templates, tool pages and explainer articles, because searchers typing it want all three. When a SERP is split, the page that wins usually serves the dominant intent and acknowledges the other in passing.

It is also worth knowing that the four-type model is a simplification. The current top-ranking article on this query argues for roughly two dozen finer-grained intents, things like "fix a technical problem", "self-diagnose symptoms", "find store locations". That granularity is useful when you are choosing a page format, and it is overkill when you are just deciding whether to write a guide or a comparison.

What our own data says about intent and clicks

Here is the part missing from every guide we read on this keyword. We classified all 12,145 queries our site received over 90 days by intent, using keyword patterns, and joined them to impressions and clicks.

Two numbers carry the whole table. Navigational queries are 14% of impressions and 91% of clicks. Everything that is not our own brand name converts at between 0.04% and 0.11%, which is roughly a hundredth of the brand rate.

Strip the brand queries out and the site earned 316 clicks from 369,997 non-brand impressions, a CTR of 0.085%.

So what does that mean for intent as a framework? Not that it is wrong. Intent still determines what kind of page you must build, and building the wrong format guarantees you do not rank at all. What the data says is that intent classification predicts what to write, not whether anyone will click. The variable that predicted clicks on our site was whether the searcher already knew who we were.

Three honest caveats before anyone quotes this. This is one site in a crowded SEO tooling niche with an informational skew, so the brand concentration is probably sharper than average. Much of that non-brand impression volume sits at poor average positions, so position is confounded with intent here. And the unclassified bucket is 61% of impressions, mostly long conversational queries that keyword patterns do not catch, so treat the four labelled rows as indicative rather than exhaustive.

Why non-brand intent earns so little: the zero-click problem

Position is not the explanation. We checked that separately, filtering to top-10 positions only so ranking is held constant, and split by query length:

• Queries of 1 to 2 words: 390 queries, 6,737 impressions, 324 clicks (4.8%)
• Queries of 4 or more words: 2,954 queries, 82,415 impressions, 35 clicks (0.04%)

2,928 of those 2,954 long-tail top-10 queries produced zero clicks. Separately, 67 of our pages sit at an average position inside the top 10 and have never been clicked.

This inverts the standard advice that long-tail queries convert better. On our data, long and specific queries in the top 10 convert a hundred times worse than short ones. The mechanism is that the answer to a specific question is now visible in the result itself, in an AI overview, a featured snippet or a people-also-ask panel, so the click is unnecessary.

That has a direct consequence for how you treat informational intent. Matching it correctly still gets you the impression and the citation. It increasingly does not get you the visit, and a content plan that assumes otherwise will look like it is failing when it is actually working as designed.

How Google itself classifies intent

The four-type model is an SEO convention, not Google's own vocabulary, and knowing Google's version is useful because it maps onto page formats more directly.

Google's Search Quality Rater Guidelines, the public document its human evaluators work from, describe user needs in four categories:

• Know queries, where someone wants information. Split further into "know simple", where a single fact answers it completely, and broader know queries that need explanation.
• Do queries, where someone wants to accomplish something: download, buy, book, use a tool.
• Website queries, where someone wants a specific site. This is navigational under another name.
• Visit-in-person queries, where someone wants a physical place nearby.

Two of those distinctions do practical work that the SEO framework hides.

"Know simple" is a warning label. If a single sentence fully answers the query, the answer will appear in the results page and you will not get the click. That is not a content problem to solve; it is a reason to weight the query lower when planning. Our zero-click numbers above are concentrated in exactly this category.

"Visit in person" has no equivalent in the four-type model at all, which is why local intent gets mishandled so often. A query with local intent needs a location page, a Google Business Profile and citations, not a blog article, and no amount of good writing substitutes.

Optimising for each intent, specifically

The table above is the summary. Here is what each one actually requires.

Informational. Answer the question in the first two sentences, then expand. Use the question wording in the heading. Add the thing that cannot be summarised: a number, a screenshot, a tested result, a stated limitation. Do not add pricing tables or signup forms, because the reader has not asked to buy and their presence weakens the page for the intent it is winning.

Navigational. Make sure your own brand pages are unambiguous and indexed: pricing, login, docs, contact, each with a title that says exactly what it is. The failure mode is a competitor or an aggregator outranking you for your own name, usually because your page titles are clever rather than literal.

Commercial. Match the entry count of the pages that rank, which is usually eight to sixteen. Give a real recommendation rather than a neutral survey, state prices, and include at least one honest limitation of the option you favour. A comparison that praises everything equally is useless and reads as sponsored.

Transactional. Remove friction rather than adding words. Price visible, what happens next stated, no required account before the value is clear. Long-form content on a transactional page usually pushes the action below the fold and costs conversions.

Mixed. Pick the dominant format from the SERP, serve it properly, and put one clear link to a page serving the other intent. Splitting a single page down the middle produces something that ranks for neither.

How to identify intent for any keyword

Do not guess from the words. Search the query and read the results page, because the SERP is Google's own answer about what the intent is.

Work through four questions in order:

1. What page types are ranking? If eight of the ten results are listicles, the intent is commercial research and a definitional guide will not rank however good it is. If they are product pages, it is transactional. If they are forum threads, people want opinions from strangers, not a brand article.

2. What SERP features are present? A featured snippet or an AI overview signals informational intent and a question with a short answer. Shopping results and ads signal transactional. A local pack signals the searcher wants somewhere nearby. Sitelinks on one dominant result usually mean navigational.

3. What do the titles promise? Read the top five titles as a set. "How to", "guide", "what is" means informational. "Best", "top 10", "vs" means commercial. Prices and product names mean transactional.

4. Do the results disagree with each other? If half are guides and half are tool pages, you have mixed intent. Decide which half you can beat, serve that, and reference the other.

Once you rank at all, Search Console is a better source than any tool, because it shows the queries you actually receive rather than the ones a database predicts. Sort by impressions, look at positions 11 to 30, and read the query wording. That is real intent data about your own pages.

For queries you do not yet rank for, question-based tools show you the shape of demand around a topic.

The trap with both is that they return questions regardless of whether the questions have search volume or whether you could ever rank for them. Use them for structure, then check the SERP for each one you plan to target.

Matching content to each intent

The most common expensive mistake is writing an informational guide for a commercial query. You research "best project management software", write 3,000 authoritative words explaining what project management software is, and rank nowhere, because every searcher wanted a ranked list with prices and every competitor gave them one.

The second most common is the reverse, and it is subtler. You have a genuinely useful explainer and you bolt a pricing table and three calls to action onto it because it gets traffic. That converts nobody, because the reader came to learn, and it weakens the page for the intent it was actually winning.

Where intent fits in a content plan

A workable sequence, given everything above:

1. Group your keywords by intent before you plan formats. One page per intent per topic, not one page per keyword. This is also the cheapest protection against cannibalisation, which is what happens when you write four pages for four keywords that share one intent. 2. Weight commercial intent higher than its search volume suggests. On our data it produced almost no clicks, but the clicks it does produce are people comparing purchases. Informational volume looks better and converts worse. 3. Expect informational pages to earn citations rather than sessions, and measure them on impressions and position rather than only on traffic. Otherwise you will kill pages that are working. 4. Do not fight navigational queries for other brands. A page targeting a competitor's name ranks below their own site and reads as opportunistic. Comparison pages that name a competitor honestly are a different thing and they do work. 5. Recheck intent every six months on your important queries. SERPs change format. A query that returned guides two years ago may return tools now, and a page that was correctly matched is now mismatched without anyone touching it.

Point 5 catches people out more than it should. Intent is a property of the results page, not of the words, so it drifts without warning.

Optimising informational intent for AI answers

If informational queries increasingly end without a click, the goal shifts from earning the visit to being the source that gets named. Three things make a page citable, and all three are within your control:

• Answer in the first two sentences under the heading. Extraction takes passages, not pages. A page that buries its answer under preamble loses to one that does not.
• Make at least one claim specific and attributable. A number with a method and a date survives summarisation. "Understanding your audience is important" does not, because a thousand pages say it.
• Phrase headings as the question. "What is search intent" rather than "Intent fundamentals". The heading is the retrieval key.

That is the reasoning behind the data section on this page, and it is worth being explicit about it: the table above exists partly because an original measurement is the one thing on a page about search intent that cannot be summarised from somewhere else.

For more on this shift, see our guide to getting AI tools to recommend your business, and the full long-tail click analysis in creating long tail keywords.

Where tooling helps, and where it does not

Intent classification is one of the few SEO tasks software genuinely does well, because it is pattern matching on query wording plus SERP composition. Distribb classifies intent as part of keyword research and groups keywords into one page per intent, which is the step that prevents you writing four competing pages on one topic. Our guide to keyword research automation covers that workflow.

Here is the limitation, and the data on this page is the evidence for it. No tool, ours included, can tell you whether an intent is worth serving on your site. Our own commercial queries carry 71,195 impressions and produced 28 clicks. A classifier looking at that keyword set would report a healthy commercial opportunity, correctly, and be silent on the fact that we were not winning it. Intent tells you what format to build. Whether you can win the format is a judgement about your own authority, and that judgement is not automatable.

The other thing to watch is that automated classification is confidently wrong on mixed-intent queries, which are exactly the ones where the decision matters most. Check the SERP yourself on any query you are planning a page around.

Frequently asked questions

What are the four types of search intent? Informational, navigational, commercial and transactional. A fifth practical category, mixed intent, covers queries whose results page shows more than one format.

How do I find the search intent of a keyword? Search it and read the top ten results. The page formats that rank are Google's own statement of the intent. Titles and SERP features confirm it.

Which intent type is most valuable? Commercial, in most cases, because those searchers are choosing between options and have not decided yet. Informational usually has the highest volume and the lowest conversion.

Does search intent affect rankings directly? Not as a ranking factor. It affects rankings by determining whether your page format matches what searchers want, and a mismatched format does not rank regardless of quality.

Can one page target more than one intent? Rarely well. Serve the dominant intent and link to a separate page for the other. Pages that try to do both usually do neither.

Why do my informational pages get impressions but no clicks? Because the answer is visible in the results page itself. On our own site, 67 pages rank in the top 10 on average with zero clicks, and long-tail queries in top-10 positions converted at 0.04% against 4.8% for short queries. This is normal now, not a fault in your pages.

How often should I recheck intent? Every six months on your commercially important queries, and after any noticeable ranking drop. SERP formats change without notice, which silently mismatches a page that was correct when you wrote it.

Does intent matter for AI search? Yes, and more than before. Informational intent is where AI answers absorb the click, so those pages should be written to be cited rather than visited: answer up front, one specific attributable claim, headings phrased as questions.

If you take one thing from this page, take the gap between the two CTR columns in the table above. Intent tells you what to build. It does not tell you that anyone will come, and knowing the difference is what stops a content plan from being judged against the wrong number. Distribb handles the classification and grouping side of that; the authority judgement stays with you.