Search intent: how to tell what Google thinks a query means

Most articles about search intent give you four categories and tell you to pick one. That is the least useful part of the subject.

The categories are easy. The hard and valuable skill is different: intent is not something you decide, it is something Google has already decided, and if your judgement disagrees with theirs, theirs wins. Every time. Your page can be excellent and still rank nowhere because it answers a question the searcher was not asking.

So the question is not "what does this keyword mean?" It is "what has Google concluded this keyword means, and can I tell?"

You can tell. The results page is where Google publishes its answer.

The categories, quickly

Worth knowing, not worth dwelling on.

| Intent | The searcher wants | Typical shape | |---|---|---| | Informational | To understand something | "how does…", "what is…", "why…" | | Commercial investigation | To decide between options | "best…", "X vs Y", "review", "alternatives" | | Transactional | To do or buy something now | "buy…", "book…", "download…", "near me" | | Navigational | To reach a specific place | A brand or product name |

Two complications that the neat table hides, and that matter more than the table:

Queries are often mixed. "Best running shoes" is commercial, but the results usually contain a mix of buying guides and shop listings, because searchers using that phrase want both. A page that is purely one or the other competes against half the results.

Intent changes over time. A query that returned articles two years ago may return product pages now, because behaviour shifted and Google followed. Intent assessed once and written down is a decaying asset.

Why guessing from the words fails

Two examples that look obvious and are not.

"Payment gateway" looks informational — someone learning what one is. Search it and the results are largely providers selling payment gateways. Google has concluded the phrase is used by people shopping, not students. Write the explainer and you are competing in the wrong category.

"Coffee grinder review" looks like it wants one review. The results are usually roundups of many grinders. Google has concluded people using that phrase want to compare, not to read about a single product.

In both cases the words support the wrong conclusion, and thirty seconds of looking would have corrected it. That is the entire method.

The method: read the page, not the keyword

Search the query and answer five questions. Ten minutes, before you write anything.

1. What TYPE of page is ranking?

Not which sites — what kind of thing they are. Articles? Product pages? Category pages? Forum threads? Videos? Tools?

If eight of the ten results are category pages on shops and you were planning a blog post, that is your answer. You can still write the post, but you are choosing to compete against a page type Google currently prefers, and you should know you are doing it.

2. What FORMAT keeps appearing?

Within the type, look at shape. Are they lists? Step-by-step guides? Comparison tables? Long explainers or short answers? If every result is a numbered list and yours is an essay, the mismatch is real even if your essay is better.

This is not an instruction to copy. It is information about what the audience has repeatedly chosen.

3. How OLD are the results?

If everything on the page was published or updated in the last few months, freshness is part of what Google thinks this query wants, and an evergreen piece will slide down over time. If the top results are four years old, freshness is not a factor and you are free to write something permanent.

4. WHO is ranking?

Are these big publishers, small specialists, retailers, forums, official sources? A results page dominated by government and medical institutions is telling you something about the trust bar for that topic. A page full of small niche blogs is telling you the opposite — that is an opening.

5. What is Google showing INSTEAD of results?

The features are direct statements of intent:

| Feature present | What it suggests | |---|---| | Map results | Local intent — a page cannot win it, a profile might | | Shopping listings | Transactional; the page wants products | | A featured snippet | There is a short factual answer, and it can be captured | | "People also ask" | The subject has adjacent questions worth covering | | Video results | The audience prefers to watch this being done | | Very few ads | Little commercial value — often lower competition, and lower reward |

There is a fuller treatment of this in how to read a search results page before you write anything.

A worked example

An illustrative case. A firm selling accounting software considers targeting "invoice template".

The words suggest transactional intent: someone wants a template. Reasonable guess. Then look:

• Type: mostly pages offering downloadable templates, plus two tools that generate one in the browser.
• Format: short pages, minimal prose, the download or the tool immediately visible.
• Age: mixed, no freshness pressure.
• Who: software companies and a couple of large template libraries.
• Features: an image block, no map, few ads.

Conclusions: the guess was directionally right but incomplete. Google thinks this query wants a thing you can take away in one click, not writing about invoicing. A 2,000-word article on invoicing best practice, however good, is answering a different question.

The correct response is either to build the actual asset — a real template or generator — or to leave the query alone and target something the firm can genuinely serve. That decision, taken in ten minutes, saves two days.

The mixed-intent case, which is most of them

Clean single-intent queries are the minority. Most commercially interesting ones are contested, and the results page shows the split.

Search a term like "project management software" and you will typically find vendor pages, a comparison article or two, and possibly a forum thread — because people typing it are at three different stages. Google hedges by serving all three.

That hedge is an opportunity, and it is where most people misread the page. The instinct is to match the majority type. The better move is to ask which slice is under-served. If eight results are vendor pages and two are comparisons, the two comparisons are absorbing all the demand from people who want to compare — and if they are mediocre, that is the gap.

A practical rule: count the result types, then ask which type you could produce the best version of. Matching the majority means competing with the strongest, most established pages. Serving a minority slice well often means competing with two weak ones.

Intent differs by who is searching, not just what

The same words can carry different intent depending on the market and the searcher, and the results reflect it.

"Insurance" typed by someone in a country with a state system and by someone in a country without one are not the same query, and the results differ accordingly. A term that is transactional in one market can be informational in another. If you sell into more than one country, the intent assessment has to be repeated per market — the same page can be correctly targeted in one and wrongly targeted in another.

The same applies to device. Queries made on a phone skew toward immediate and local intent more than the same words typed on a desktop, which is part of why map results and "near me" behaviour dominate mobile pages.

When your page and the intent disagree

Three options, and it helps to name them.

Change the page. Usually right. If the query wants a comparison and you wrote an explainer, write the comparison.

Change the query. Also often right. Find the phrasing that matches what you actually have. A firm with genuine expertise in invoicing rules should target the questions where that expertise is what the searcher wants.

Compete anyway. Occasionally correct — when you can see the results are poor and you have something genuinely better. Rare, and it should be a decision rather than an oversight. Going against established intent needs a strong reason, not a strong opinion.

How this connects to everything else

Intent decides what you write, which decides whether the work counts at all. It sits upstream of nearly everything:

• It determines which keywords are worth targeting, which is why it belongs inside keyword research rather than after it.
• It explains a large share of "why is my page not ranking?" — see the diagnostic checklist.
• It is one of the inputs described in how Google ranking works.

It also sets a limit worth being honest about. If a query's results are entirely marketplaces, directories and large retailers, a small site's page is unlikely to break in regardless of quality — the intent is being served by a page type you are not. In that situation the realistic routes are a different query, or appearing on one of those platforms rather than beside them. The trade-offs are in when renting attention beats owning it, and the actual prices are on the rankings board.

Common questions

Can one page serve two intents? Sometimes, when the results themselves are mixed. Lead with the dominant one and satisfy the second below.

How often should I recheck intent? For pages that matter, once or twice a year, and immediately if a page loses position without an obvious cause.

Does matching intent guarantee ranking? No. It is necessary, not sufficient. Mismatching it is close to sufficient for failure, which is why it is worth ten minutes.

Is intent the same in every country? No. The same words can return different result types in different markets. Check the market you sell to.