Keyword Research: How to Find High-Intent Keywords

Keyword research is the process of finding the words and questions people type into search engines, then judging which of those searches are worth building a page around. Done well, it is the difference between publishing content and publishing content someone was already looking for. This guide is for founders, marketers, and writers who want a repeatable method rather than a folder full of exported spreadsheets. It covers search intent, keyword types, how to build and expand a seed list, clustering, SERP analysis, entity research, and a simple way to decide what to work on first. The thread running through all of it is one idea: intent matters more than volume.

The short version

• The highest-volume keyword is rarely the one worth ranking for. What a searcher wants, and whether you can serve it, decides the value.
• There are four search intents: informational, commercial, transactional, and navigational. Read them before you write anything.
• One page should target a cluster of related keywords, not a single phrase. One-keyword-per-page thinking causes more cannibalization than it prevents.
• The SERP is your most honest keyword tool. The pages Google already ranks tell you what it thinks the query means.
• Prioritize with relevance, intent, difficulty, and business value together. Volume alone is a trap.

What is keyword research, and why does intent beat volume?

Keyword research is intent research with a keyword attached. The keyword is the label; the intent is the reason someone searched. A term with heavy search demand is worthless to you if the people typing it want something you don't sell, and a quieter term can be your best page if it catches someone ready to act.

Volume tells you how many people search. It says nothing about whether they will convert, whether you can realistically rank, or whether the click helps your business. Those three questions decide value, and none of them appear in a volume column.

This is also why volume-first research ages badly. You chase a big number, publish a page that half-matches the intent, and it stalls on page two while a smaller, sharper page outranks it. If you want the wider context for where research sits in a program, our complete guide to SEO maps the full sequence from research to measurement.

What are the four types of search intent?

There are four, and every keyword you consider belongs to one of them. Naming the intent first stops you from writing the wrong kind of page.

• Informational. The searcher wants to learn. Queries like how to, what is, and examples of. They want an answer, not a checkout.
• Commercial. The searcher is comparing before buying. Modifiers like best, vs, alternatives, and review. They are close, but still deciding.
• Transactional. The searcher wants to act now: buy, sign up, book, download. Terms like pricing, buy, demo, and branded product queries.
• Navigational. The searcher wants a specific site or page and is using search as a shortcut, for example notion login. Usually only worth targeting for your own brand terms.

A single word can shift intent completely, and each of these four demands a different kind of page. Reading the wrong one is the most expensive mistake in keyword research, which is why our guide to search intent covers how to identify each type, which modifiers are genuinely reliable signals, and how to spot a mismatch on a page you have already published.

How do you read intent from the SERP?

Search the keyword and study what already ranks. Google has spent years learning what satisfies a query, and the results are its published answer. If the first page is full of listicles, that is the format it rewards; if it is product pages, an article will struggle no matter how good it is.

The dominant content type, the page titles, and the SERP features together tell you what Google thinks the query means. We break down how to read each of those three signals, and what to do when the first page mixes formats, in the search intent guide. For how Google forms these results in the first place, see our explainer on how Google Search works and Google's own documentation on how Search works.

What are the main keyword types, and when does each matter?

Keywords fall into a small number of shapes, and each earns its place at a different point. The table below is the working version.

Keyword type	What it looks like	When it earns a place
Head	One or two words, broad: crm, email marketing	Rarely as a direct target. Use it as a topic anchor, not a page.
Body	Two to three words, more specific: crm for startups	When you have some authority and the intent is clear enough to satisfy.
Long-tail	Longer, specific phrases: best crm for a two-person sales team	Early on, and always. Lower competition, sharper intent, better conversion.
Question	Explicit questions: how do i migrate crm data	For informational pages, FAQs, and AI-answer capture.
Entity / topic	Concepts and named things: a product, method, or category	For building topical depth so a whole cluster ranks, not one page.

Head terms flatter your ego and drain your calendar. Long-tail terms win now and compound into the head term later, because a site that covers the specifics earns the authority to rank for the broad phrase.

How do you build a seed list?

Start from what you already know, not from a tool. Write down the words your customers use for the problem you solve, the words your product pages use, and the questions you hear in sales calls and support tickets. That is your seed list, and it is more honest than any export because it comes from real demand.

Then widen it with a few free sources. Google autocomplete shows real query patterns as you type. The People Also Ask and related searches on the SERP hand you adjacent questions. Your own site search and Search Console queries show what people already arrive on you looking for.

Twenty to forty good seeds are enough to start. You are not trying to be exhaustive yet, just to map the territory.

How do you expand a seed list into a full keyword set?

Take each seed and pull on it in three directions: modifiers, questions, and adjacent entities. Add commercial modifiers (best, vs, for [audience]), add the natural questions around it, and add the related concepts a thorough page would mention.

Tools speed this up. Keyword tools such as Ahrefs, Semrush, or Google Keyword Planner will suggest matching terms, related terms, and questions for any seed. Treat their metrics as directional estimates from a tool, not as facts. Different tools disagree on the same keyword because they model demand differently, so use them to find candidates and rank them roughly, then trust the live SERP for the final read on intent and difficulty.

Expansion is where lists balloon. Resist the urge to keep everything. A keyword you can neither satisfy nor benefit from is not an opportunity, it is a distraction.

What is keyword clustering, and why group keywords into topics?

Clustering is grouping keywords that share the same intent so one page can target the whole group. If several phrases return roughly the same ranking pages, Google treats them as the same query, and they belong on one page, not five thin ones.

Group by intent and by SERP overlap, not by wording. how to reduce cart abandonment and ways to stop customers abandoning carts are one cluster. cart abandonment software is a different one, because the intent flips from learning to buying.

Clustering is the antidote to two problems at once. It stops you spreading effort across near-duplicate pages, and it tells you how deep a single page needs to go to satisfy the full group. Our guide to content marketing for SEO covers how those clusters become a content calendar.

What can SERP analysis tell you about a keyword?

More than any metric. Before committing to a keyword, read the first page closely and answer four questions.

• What content type ranks? Articles, product pages, tools, or videos. That is the format you have to match or beat.
• Who ranks? If the page is filled with sites far larger and older than yours, the term may be out of reach for now. A full competitor analysis tells you whether that gap is closing or widening. Note it and move on.
• How good is the current best result? If the top pages are thin or dated, there is room. If they are excellent, you need a genuine reason to exist.
• What is the sub-intent? The titles reveal the specific angle searchers reward, which is often narrower than the keyword suggests.

This is also your difficulty read. Keyword-difficulty scores from tools are a starting estimate; the actual competition is whoever is on page one right now.

What is entity and semantic research?

Entities are the concepts, people, products, and categories a topic is built from, and covering them is how you show depth rather than keyword repetition. Search engines increasingly understand topics as networks of related entities, not bags of matching words, so a page that mentions the concepts a real expert would mention reads as more complete.

Find them the same way you find keywords. The related searches, the People Also Ask questions, the recurring subheadings across ranking pages, and the terms your competitors keep using all point to the entities a thorough page should cover. Google's Search Essentials frames this as making genuinely helpful, comprehensive content rather than gaming a term.

Semantic research does not replace keyword research. It decides what goes inside the page once the keyword has decided the page exists.

How do you prioritize keywords? The RIDB rule

Score every candidate on four factors and let the math sort them, so effort follows opportunity instead of enthusiasm. We call it the RIDB rule: Relevance, Intent match, Difficulty, and Business value. Rate each from 1 to 3, then apply one formula.

Priority score = Relevance + Intent match + Business value − Difficulty.

• Relevance (1–3): how closely the topic sits to what you actually do.
• Intent match (1–3): how well you can satisfy what the searcher wants with the page and offer you have.
• Business value (1–3): how much a ranking here moves revenue, leads, or authority.
• Difficulty (1–3): how hard page one looks, read from the live SERP. This one is subtracted.

Scores range from −2 to 8. Work the highest first. Here is a worked example using illustrative ratings, not real search volumes, for a hypothetical project-management tool.

Example keyword	R / I / B / D	Score and read
how to write a project plan	3 / 2 / 2 / 1	6: informational, on-topic, winnable now. Write it.
best project management software	3 / 3 / 3 / 3	6: high value but hard. A long-term target, not this month.
asana pricing	1 / 3 / 1 / 1	4: clear intent, but a competitor's brand term. Low fit.
project management	2 / 1 / 2 / 3	2: broad and ambiguous. Deprioritize.

Two keywords tie at 6, and the tie is the point: one is easy and modest, one is valuable and hard. Ship the winnable one now, and let it build the authority you will need for the hard one later.

When not to chase a keyword. Skip it when your relevance is low, however big the term looks, because off-topic traffic does not convert and dilutes what your site is about. Skip it when the SERP is settled and owned by sites far beyond your current authority. And skip it when the intent is transactional but you have nothing to transact. A high score on a keyword you can't serve is still a no.

How do you map keywords to funnel stage and content type?

Match the intent to the stage, then the stage to the format. Informational sits at the top, commercial in the middle, transactional at the bottom, and the content type follows from there.

Funnel stage	Dominant intent	Content type it maps to
TOFU (awareness)	Informational	Guides, explainers, how-to articles, glossaries
MOFU (consideration)	Commercial	Comparisons, “best” lists, alternatives, reviews
BOFU (decision)	Transactional	Product and service pages, pricing, demos, case studies

The mistake here is answering a top-of-funnel question on a bottom-of-funnel page, or the reverse. Someone searching what is an employer of record is not ready for a pricing page, and someone searching eor pricing does not want a definition. Serve the stage they are in. When the keyword is transactional and matches a service, the right destination is a page like our SEO services page, or a dedicated landing page built for that exact query, which is where our landing page SEO playbook picks up, not another article.

What are the most common keyword research mistakes?

Five recur often enough to name.

• Chasing volume. Picking the biggest number and ignoring whether the traffic converts or whether you can rank at all.
• Ignoring intent. Writing an article for a query that wants a product page, then wondering why it never ranks.
• One-keyword-per-page thinking. Building a separate thin page for every phrase instead of one strong page per cluster.
• Cannibalization. Two of your own pages targeting the same intent, so they split signals and neither wins.
• Overreaching on authority. Targeting terms owned by far larger sites while easier, more valuable long-tail terms sit untouched.

Once the keywords are chosen, the page still has to be built to serve them. Our on-page SEO checklist covers turning a keyword cluster into a page that satisfies it, and Google's Search Central documentation is the reference for the technical side.

The habit worth building is simple: never open a document before you have read the SERP for the keyword. The results page is the brief, and skipping it is how good writing ends up on the wrong kind of page. This week, take your ten most important keywords, run each through the RIDB rule, and cluster the survivors into topics. You will usually find that ten keywords collapse into three or four pages worth writing.