Keyword relevance is how well a page answers what a searcher meant by one specific query. Google judges it per query, not per page, by reading the words on the page, the meaning behind them, whether the format fits what the searcher expects, the links pointing in and how people respond to the result. Placing the keyword in the title and H1 is the smallest part of it: a page can repeat a phrase twenty times and still be irrelevant to it. This guide covers the three layers Google checks, where it reads the signals, and how to see which queries Google already thinks your page is about.
What Keyword Relevance Actually Means
Relevance is not a property your page owns; it is a match between one page and one query. Your guide to “running shoes for flat feet” may be highly relevant to that query, partly relevant to “best stability running shoes” and irrelevant to “flat feet exercises,” even though all three share words.
Google describes the starting point plainly in its How Search Works pages: “The most basic signal that information is relevant is when content contains the same keywords as your search query,” and if those keywords appear “in the headings or body text of a webpage, the information might be more relevant.” Note the word “basic.” The same page explains that Google builds language models to work out what a short query means and uses a synonym system to find relevant documents that do not contain the exact words.
Relevance and Quality Answer Different Questions
Think of a two-step filter. Relevance decides whether your page is a candidate for the query at all; quality, authority and usefulness decide where it sits among the candidates. Google’s ranking page lists meaning, relevance, quality, usability and context as separate inputs.
That explains why a strong site can publish a page that never ranks (wrong intent, wrong format), while a thin page on a weak site ranks for a long-tail query because almost nothing else is relevant to it. When a page is stuck, diagnose which filter it fails before rewriting. Links rarely rescue a page that fails the relevance test, and rewriting a relevant page that simply lacks authority rarely helps either.
The Three Layers Google Checks
Split relevance into three layers; each fails in a different way.
| Layer | The question it answers | What it looks at | How it fails |
|---|---|---|---|
| Lexical match | Does the page use the words of the query? | Title, headings, body text, anchor text | The page never names the thing the searcher typed, or buries it |
| Semantic relevance | Does the page cover the concept, including synonyms and related entities? | Language understanding across the whole document and its passages | The page mentions the term but does not explain it, or covers a neighbouring topic |
| Search intent match | Is this the kind of page the searcher wants? | Page type, format, angle and depth compared with what satisfies searchers | A blog post for a “buy” query, a product page for a “how to” query |
Old-school SEO lived in the first row; competitive queries are now decided in the second and third. When Google introduced BERT in October 2019, it said the model helped Search understand “one in 10 searches in the U.S. in English,” with examples where small words like “to” and “for” changed what the query meant (our BERT explainer covers what changed). The lexical layer still matters, but mainly as a clear label.
Where Google Reads Relevance on a Page
Some elements carry far more relevance signal than others, and a few carry none. Google’s SEO Starter Guide settles several of the old debates directly.
- Title and main heading. The clearest label of what the page is about.
- Body text. Where semantic relevance is earned: definitions, steps, examples, related concepts.
- Anchor text of links pointing to the page. Google’s link documentation says anchor text “tells people and Google something about the page you’re linking to,” and that “the words before and after links matter.” Your own internal links are the one anchor source you fully control.
- Not the meta keywords tag. The Starter Guide states that “Google Search doesn’t use the keywords meta tag.”
- Barely the domain or URL. Keywords in the domain or URL path “have hardly any effect beyond appearing in breadcrumbs,” per the same guide.
- Not length. “The length of the content alone doesn’t matter for ranking purposes.” A longer page is only more relevant if the extra words answer more of what the searcher needs.
Repetition is the other trap. The Starter Guide warns that “excessively repeating the same words over and over (even in variations) is tiring for users,” and Google’s spam policies define keyword stuffing as filling a page with keywords “in an attempt to manipulate rankings.” There is no target density; our piece on keyword density explains why the old percentages were never a real signal.
Anchors, Body, Clicks: What the Antitrust Case Showed
The US Department of Justice’s search antitrust case added a rare look inside. Notes from a 2025 call with a Google ranking engineer, published by the DOJ as a trial exhibit, describe “ABC” signals (anchors, body and clicks) as the key components of “topicality,” characterised as Google’s base relevance measure. Anchors are links from other pages, body is the terms in the document, and clicks reflect how long users stayed on a result before returning to the results page.
Treat these as notes summarising one conversation, not a full specification of a ranking system that changes constantly. They line up with what Google says publicly, including that it uses “aggregated and anonymized interaction data to assess whether search results are relevant to queries.” The takeaway: relevance is not only something you write. It is also what other pages say about you, and what searchers confirm or reject.
Search Intent Match: The Layer Keyword Tuning Cannot Fix
Search intent match is where most “relevant” pages actually lose. If the top results for a query are all category pages and you publish a long guide, no keyword placement makes the guide the kind of page searchers want.
Read the results page before you write: the dominant page type (guide, list, product, category, tool), the angle (beginner, comparison, cheapest) and the SERP features. Our guide to search intent walks through the four classic intent types and the mixed cases, and the free Search Intent Checker gives a fast first read. Inside the SEO Rocket app, keyword research runs on real Ahrefs data and returns intent alongside volume and difficulty, so a query whose intent your page cannot serve gets dropped before anyone writes a word.
Topical Relevance: Page Level and Site Level
Topical relevance works at two scales. At page level, it is whether the document covers the subject the way a knowledgeable person would: the sub-questions, definitions and related entities. A page that names a topic often but never explains its parts reads as thin.
At site level, it is whether the rest of your site gives the page context. A payroll compliance article on a wedding photography site has nothing around it: no supporting pages, no related internal links, no history of being linked to for that subject. Clusters of pages on one subject, linked with descriptive anchors, make that context explicit. Our guides to topical authority and semantic SEO cover the planning side. Be honest about the limit: site-level topical relevance helps a relevant page compete; it does not make an off-intent page relevant.
How to See Which Queries Google Thinks Your Page Is About
The best evidence of how Google reads your page is the list of queries it already shows the page for. Most guides skip this, and it is free. In Google Search Console, open the Performance report, filter by the page URL, and switch to the Queries tab. That list is Google’s own view of your page’s topic, and comparing it with the topic you intended is the fastest relevance audit there is.
Work through it in this order:
- Queries you intended, ranking low. Relevance is probably fine; the gap is depth, authority or freshness.
- Queries you did not intend, with real impressions. Google sees a different or broader topic than you planned. Either the page is drifting (sections that belong on another URL) or there is demand you should give its own page.
- Your target query missing entirely. The page fails the first filter: the term is absent or buried, or the intent does not match.
- Another of your URLs showing for the same query. Two pages are splitting the signal. That is keyword cannibalization, and the fix is consolidation or a clearer split of jobs.
A worked example: say a page titled “Water Heater Installation Guide” shows impressions mainly for repair and troubleshooting queries, with the installation query nowhere near page one. That page is not under-optimised for “installation.” It is over-relevant to repair, probably because its longest sections are about fixing faults. Move the repair material to its own page, link the two with descriptive anchors, and tighten the installation page around its one job. With Search Console connected to SEO Rocket, you can ask for this query-by-page view in chat instead of clicking through filters.
Common Relevance Mistakes
- Chasing “semantic keywords” lists. Pasting every term a content tool suggests can widen a page until it is vaguely relevant to many queries and strongly relevant to none. Add a related term only when it answers part of the searcher’s question. The LSI keywords myth grew out of this habit.
- Targeting several intents on one URL. “What is X,” “X pricing” and “X vs Y” usually deserve separate pages. Group keywords by shared results, not shared words; the free Keyword Grouping Tool helps here.
- Generic internal anchors. “Click here” and “read more” waste the relevance signal you control most directly.
- Writing to a word count. Length alone does not matter; depth on the right sub-questions does.
A Pre-Publish Relevance Checklist
- The page type and angle match what currently ranks for the target query.
- The title and H1 name the topic in the words searchers use, once and clearly.
- The first paragraph answers the core question directly.
- Headings cover the sub-questions a searcher has, not padding sections.
- Related concepts and entities appear where an expert would naturally use them.
- No other URL on the site targets the same intent (check your keyword map).
- Related pages link in with descriptive anchor text.
FAQ
Is keyword relevance a ranking factor?
It is better thought of as the entry condition for ranking. Google’s documentation names relevance as one of the main things it assesses, but relevance on its own only makes a page a candidate; quality and usefulness decide its position.
Do I need the exact keyword on the page?
Not always; Google’s language systems understand synonyms and paraphrases. The searcher’s wording in the title and main heading is still the clearest label, so use it where it reads naturally.
How is it different from topical relevance?
Keyword relevance is the match between one page and one query. Topical relevance is broader: how thoroughly a page, and the site around it, covers a subject. Strong topical relevance makes it easier for each page to be judged relevant to many related queries.
Does Google Ads use the same idea?
Google Ads has its own version: Quality Score combines expected click-through rate, ad relevance and landing page experience. The idea is similar, but it is a separate system from organic ranking.
The Short Version
Keyword relevance is not a density target or a placement rule. It is Google’s per-query judgment that your page uses the searcher’s language, covers the concept and is the kind of page that satisfies the search. Get the intent right first, label the page clearly, cover the topic like an expert, then check the Search Console query list to confirm Google reads the page the way you meant. That order of work sits behind a playbook proven across 1,000,000+ ranking pages, and SEO Rocket’s research, writing and audit steps follow it.
