Keyword Research & Search Intent

Keyword research is the process of discovering the words and phrases people actually type (or speak) into search engines, then evaluating those phrases so you can decide which ones are worth targeting. It sits at the very start of any serious SEO project because it connects two things you cannot control by guessing: real demand and real language. If you build pages around terms nobody searches for, or around terms so competitive you have no chance of ranking, you will pour effort into content that never earns traffic. Done well, keyword research turns SEO from a hopeful exercise into a demand-led plan.

The goal is not to collect a giant list of words. It is to understand what your audience is trying to accomplish, in what language, and at what stage of their journey. That understanding shapes everything downstream: which pages you create, how you structure them, what you say, and how you measure success.

Why keyword research matters

Search engines match pages to queries. If your content does not reflect how people phrase their problems, it will not surface for those problems, no matter how good the writing is. Keyword research gives you three concrete advantages. First, it reveals demand — you learn which topics have an audience before you invest in creating content. Second, it reveals intent — the underlying reason behind a search, which tells you what kind of page will satisfy it. Third, it reveals opportunity — the gap between what people want and what the current results actually deliver, which is where new pages can win.

Research also protects you from wasted effort. A single well-chosen topic that matches a page you can realistically rank is worth more than a dozen pages chasing terms dominated by established competitors. And because keyword data ages, revisiting it periodically keeps your content aligned with how language and demand shift over time.

Finding seed keywords and expanding them

Every keyword list starts with seed keywords: the broad, obvious terms that describe your topic, product, or service. If you sell running shoes, your seeds might be "running shoes", "trail running", and "marathon training". Seeds are not your final targets — they are the doorway into the wider set of phrases people actually use.

Good sources for seeds include your own product and service names, the language customers use in support tickets and reviews, the labels competitors use, and the questions people ask in your niche. Once you have a handful of seeds, you expand them into a much larger universe of related queries. Common expansion techniques include:

• Autocomplete and related searches: the suggestions search engines show as you type, and the "people also search for" and "related searches" blocks on the results page, are direct signals of real demand.
• People Also Ask boxes: the expandable question boxes surface adjacent questions that often make excellent subheadings or standalone pages.
• Keyword tools: dedicated tools take a seed and return hundreds of variations, questions, and modifiers, each with estimated metrics you can compare.
• Competitor pages: the terms competitors already rank for reveal proven topics you may be missing.
• Modifiers: adding words like "best", "how to", "vs", "near me", "for beginners", or a year expands a seed into distinct, more specific intents.

The output of this stage is messy on purpose — a large, unfiltered list. Filtering and prioritising come next, once you can attach metrics and intent to each phrase.

Reading keyword metrics: volume, difficulty and CPC

Keyword tools attach numbers to each phrase, and three matter most. Search volume estimates how many times a term is searched in a given period, usually per month. It signals potential demand, but it is an estimate, it is often averaged across a year, and it can hide seasonality. Treat volume as a directional guide, not a precise forecast.

Keyword difficulty is a tool-specific score that estimates how hard it will be to rank on the first page, based largely on the strength of the sites already ranking. Difficulty scores are not standardised between tools, so a "40" in one tool is not the same as a "40" in another. Use difficulty to compare terms within a single dataset, not as an absolute law.

Cost per click (CPC) is the average price advertisers pay for a click on that term in paid search. On its own it is a paid-ads figure, but in keyword research it serves as a useful proxy for commercial value: advertisers only bid high on terms that convert into revenue, so a high CPC often flags a query with strong buying intent even if its volume is modest.

The critical discipline is to read these metrics together, never in isolation. A high-volume term with high difficulty may be unwinnable for a young site. A low-volume term with high CPC and low difficulty can be far more valuable, because it attracts fewer but more commercially motivated visitors against weaker competition. Volume tells you how big the prize is; difficulty tells you how likely you are to claim it; CPC hints at how much each visitor is worth. Decisions come from the combination, weighed against your own site's authority — which you build through strong content and E-E-A-T and earned links.

The four types of search intent

Behind every query is a goal. Matching your page to that goal — the search intent — is arguably more important than any single metric, because search engines rank the pages that best satisfy the intent, not the pages that merely contain the keyword. Intent is commonly grouped into four types:

• Informational: the searcher wants to learn something — "how does keyword research work", "what is search intent". They expect guides, explanations, and answers, not a sales pitch.
• Navigational: the searcher wants a specific site or page — "gmail login", "nike returns policy". They already know where they want to go and are using search to get there.
• Commercial investigation: the searcher is comparing options before buying — "best running shoes for beginners", "tool A vs tool B", "product reviews". They are close to a decision but still evaluating.
• Transactional: the searcher is ready to act — "buy running shoes online", "sign up", "cheap flights to lisbon". They want to complete a purchase or conversion now.

How to read intent from the SERP

You do not have to guess intent — the search results page tells you. The pages that already rank are, in effect, the search engine's answer to "what does this query mean". If the top results are long explanatory articles, the intent is informational and a product page will struggle there. If they are category and product pages, the intent is transactional. If they are comparison articles and "best of" lists, the intent is commercial investigation. Look also at the SERP features: shopping carousels signal transactional intent, while a featured snippet or People Also Ask block signals informational intent. Always sanity-check your assumptions against a live search before committing a page to a keyword.

Long-tail keywords and why they convert

Long-tail keywords are longer, more specific phrases — "waterproof trail running shoes for wide feet" rather than "running shoes". They usually carry lower search volume individually, but they have three advantages that make them disproportionately valuable, especially for newer or smaller sites.

• Lower competition: fewer sites target the exact phrasing, so ranking is more achievable without a large authority head start.
• Clearer intent: the added specificity reveals exactly what the searcher wants, which makes it easier to build a page that satisfies them completely.
• Higher conversion: someone searching a precise, detailed phrase is usually further along in their decision, so the traffic tends to convert better than broad, vague queries.

Collectively, long-tail queries also make up the majority of all searches, because language is endlessly varied. A content strategy that captures many specific long-tail phrases can attract more qualified traffic in aggregate than a handful of fiercely contested head terms.

Grouping keywords into topic clusters

Once you have a filtered, intent-labelled list, the mistake to avoid is building one page per keyword. Many keywords are simply different phrasings of the same underlying need — "keyword research", "how to do keyword research", and "keyword research process" all describe one topic. Search engines understand this synonymy, so a single strong page can rank for all of them.

The solution is to group keywords into topic clusters: sets of related queries that share an intent and belong together on one page or within one tightly linked group of pages. A cluster typically has a broad pillar topic supported by narrower cluster subtopics, all interlinked. This structure signals topical depth to search engines and gives readers a coherent path through the subject. Grouping is where a raw keyword list becomes an actual site architecture.

One primary intent per page: avoiding keyword cannibalization

Each page should target one primary intent and one keyword cluster. When two or more pages on your own site chase the same keyword and intent, they compete with each other — a problem called keyword cannibalization. The search engine cannot tell which page is the definitive answer, so it may rank the weaker one, split ranking signals between them, or rotate between them, and none reaches its potential.

To prevent this, map every target cluster to a single, clearly designated page before you write. If you discover overlapping pages later, the fix is usually to consolidate them into one stronger page, or to differentiate them so each serves a distinct intent. Getting this mapping right is a core part of on-page SEO, where each page's title, headings, and body are aligned to its one chosen intent.

Competitor and keyword-gap analysis

Your competitors have already done a version of this research, and their results are visible. A keyword-gap analysis compares the terms competitors rank for against the terms you rank for, surfacing valuable queries where they appear and you do not. Those gaps are shortlists of proven, achievable opportunities — topics with demonstrated demand that you have not yet addressed.

Effective competitor analysis goes beyond copying a list. Look at which competitors rank for a term (a term dominated by huge brands may be out of reach, while one held by sites like yours is winnable), what type of page ranks (which confirms intent), and where the existing results are weak — thin content, outdated information, poor user experience — because those weaknesses are your opening. The aim is to find terms you can realistically win, not just terms that exist.

Turning research into a content plan

Research only pays off when it becomes a plan you can execute. Convert your clustered, prioritised keywords into a content roadmap by working through a repeatable sequence:

• Assign one primary keyword and cluster to each planned page, with the confirmed intent noted alongside it.
• Prioritise by opportunity, balancing volume, difficulty, commercial value, and how well the topic fits your site's existing authority — start where achievability and value overlap.
• Choose the right page type for the intent: a guide for informational, a comparison for commercial investigation, a product or landing page for transactional.
• Note the subtopics and questions each page must cover, drawn from People Also Ask, related searches, and the cluster's supporting keywords.
• Plan internal links between pillar and cluster pages so the structure reinforces topical relationships.
• Schedule and revisit, since demand, competition, and language change — keyword research is a recurring cycle, not a one-time task.

With that plan in hand, the next steps are executing each page well and earning the authority to rank it. Explore how to structure and optimise individual pages in the On-Page SEO guide, how to build the trust signals that support rankings in Content & E-E-A-T, or return to the SEO guide hub to see how keyword research fits the wider picture.