How to Look for Keywords on a Website

Two of these steps deserve a closer look. Branded versus non-branded matters because the split answers different questions: branded rankings measure how well-known a site is, while non-branded rankings measure how good its SEO actually is. And the cross-reference in step four is a discipline of its own, called keyword gap analysis; the keyword gap analysis guide walks through it with worked examples.

The whole loop, from lookup to comparison to prioritized list to a page brief for each winner, is exactly what Orchory automates end to end. It researches the keywords, clusters them into page-level topics, ranks the opportunities, and turns each one into a build-ready brief, so the manual method above becomes a review step instead of a week of spreadsheet work.

There are two ways to see keywords on a website: infer them from what the page says, or measure them from what the site actually ranks for. Inference is the old manual trick. Open a page, read the title tag, the meta description, the H1 and the subheadings, and you can usually tell which phrase the author was targeting. Right-click, choose "view page source", and the <title> and meta tags spell it out even when the design hides them.

That trick is worth knowing, and worth keeping in perspective. On-page signals only reveal what a site is trying to rank for, never what it does rank for, and those two lists disagree far more often than intuition suggests: 96.55% of pages get zero traffic from Google, per Ahrefs' 2023 study of 14 billion pages. A title tag tells you the author's intent. It cannot tell you whether Google agreed, which position the page earned, or how much search volume the term carries.

Ranking data answers all three at once. A lookup tool built on real search performance data, like the checker at the top of this page, shows the keywords a site holds actual positions for, including the ones its authors never deliberately targeted. Sites routinely rank for phrasings they never wrote into a title, and just as routinely miss rankings for the exact phrases they optimized hardest for.

So use both, in the right order: pull the ranking data first, and treat view-source archaeology as a supplement. When a competitor ranks #3 for a term you care about, that is when reading their title, headers, and page structure pays off, because now you are studying a page Google has already graded, not guessing at one.

A keyword lookup gives you raw material. A keyword strategy for SEO is what you do with it, and the single biggest upgrade is to stop thinking in individual keywords and start thinking in topic clusters. Google ranks pages, not keywords, and one well-built page can rank for dozens of related phrasings. So instead of assigning one keyword per page, group related terms into clusters and build one page per cluster. If clustering is new to you, start with what keyword clustering is and how it works.

Next, map intent to page type. Informational searches ("how to look for keywords on a website") want a guide; commercial ones ("best keyword research tool") want a comparison; transactional ones ("keyword checker free") want the tool itself. A page that answers the wrong intent loses to a weaker page that answers the right one, no matter how well it is written.

Then prioritize clusters, not keywords. Sum the volume across each cluster, average its difficulty, and rank the clusters by what they could realistically return. A cluster of eight long-tail terms often beats one prestige head term: long-tail phrases are individually small, and collectively enormous, with 94.74% of keywords getting ten or fewer monthly searches, per Ahrefs' 2026 statistics roundup. The best keywords for your website are usually the specific ones your size lets you win now, not the giant ones your competitors already own.

Treat the result as a loop rather than a document. Rankings shift, competitors publish, and a cluster you passed on in spring can be winnable by autumn, so re-run the lookup on your own domain and your closest rivals every month or so and re-prioritize whatever moved. The refresh takes minutes with the checker above, and it is the difference between a keyword strategy for SEO that tracks the market and one that fossilizes in a spreadsheet.

“If you publish a page on a topic that no one is searching for, that article won’t receive any traffic from Google (or other search engines) — no matter how hard you try.”

Tim Soulo, CMO at Ahrefs, in Ahrefs' keyword research guide

The output of all this is your list of SEO strategy keywords, ranked as page opportunities: which topics deserve a page, in which order, and what each page must cover. That list is the bridge between research and results, and producing it automatically is exactly where Orchory picks up. Need more raw candidates first? The free AI keyword research tool expands any seed topic into an intent-classified keyword map you can cluster from.

Once a cluster has a page, keyword placement is simple and mostly front-loaded. Put the primary keyword in the title tag, in the H1, and somewhere in the first paragraph, phrased the way a real searcher would type it. Work the strongest supporting phrases into subheadings where they genuinely describe the section beneath them. That is the entire load-bearing part of SEO keyword usage; everything after it is writing.

In the body, write the way your searchers phrase things rather than the way your industry does. People search "how to see keywords on a website", not "website keyword visibility methodology", and matching their words in your copy is both better SEO and better writing. Use the related phrases from your cluster (the variations, the questions, the synonyms) instead of repeating the exact-match keyword, because coverage of the topic is the signal, and repetition is not.

A few placements beyond the body still earn their keep. Put the primary keyword in the URL slug and keep the slug short. Write it into the meta description too: that text does not influence ranking directly, but a description that mirrors the query wins more clicks from the results page. Give images alt text that genuinely describes them, keyword included only when it belongs. And when other pages on your site link to the page, make the anchor text say what the page is about rather than "click here". Each placement is small on its own; together they tell search engines, consistently, what the page is for.

What you should not do is stuff. Counting occurrences and chasing a keyword density percentage is a relic of a search engine that no longer exists, and Google has said so plainly for over a decade:

“Keyword density, in general, is something I wouldn’t focus on. Make sure your content is written in a natural way.”

John Mueller, Search Advocate at Google, in a Search Central office-hours, quoted in Ahrefs' SEO glossary

Modern search systems evaluate language the way readers do, and natural phrasing with broad topic coverage consistently outperforms mechanical repetition. The practical test is blunt: read the paragraph aloud, and if a sentence exists only because a keyword had to live somewhere, cut it. A page that covers SEO keywords and phrases naturally, in the searcher's own vocabulary, places every keyword exactly where it needs to be without anyone counting anything.

The outdated playbook picked a keyword, wrote a page around it, repeated the phrase until a density checker went green, and published. The best keyword strategy today inverts almost every part of that: cluster keywords by topic, map the intent across the whole cluster, and judge every opportunity by realistic traffic potential and difficulty rather than raw volume. One page targets one topic, which usually means a primary term plus five to twenty related phrasings, never a single keyword and never a stuffed list.

Realism matters more than it used to, because ranking is no longer the whole game. 68.01% of US Google searches in early 2026 ended without a click, per SparkToro's analysis of Similarweb clickstream data, so a modern strategy weighs which positions still earn visits. Top spots remain disproportionately valuable: the #1 organic result averages a 27.6% click-through rate, ten times that of position #10, per Backlinko's analysis of 4 million search results (updated 2025). A cluster you can take to the top three beats a bigger cluster where you would stall on page two.

A worked example makes that trade-off concrete. Suppose the lookup hands you two options: a head term at 27,000 searches a month with difficulty 89, or a cluster of eight long-tail questions averaging 700 searches each at difficulty 22. The head term looks five times bigger, but the realistic outcome for a smaller site is page two, which those click-through numbers say earns almost nothing. The cluster's 5,600 pooled searches, with a genuine shot at the top three, will deliver more visitors and more buyers. Multiply that judgment across every cluster in your list and you have a strategy. Chase the big number instead and you have a trophy list.

And then the part most strategies skip: shipping. Most teams do not fail at keyword research; they fail between the keyword list and the published page, where briefs, approvals, and developer queues live. A spreadsheet of perfectly clustered keywords earns nothing while it waits. The best keyword strategy in 2026 and beyond is the one that closes the loop, where every prioritized cluster becomes a live page in days, not quarters, because execution, not research, is the real bottleneck. That is the gap Orchory was built to close.