How to Do Keyword Research for SEO

Keywords Basics

Keywords, what are they and How to Do Keyword Research for SEO.

A keyword is a word or phrase that people use when searching for information on the internet through search engines.

Keyword research means understanding the language which your target audience uses when searching on the internet for information, products or any type of content.

The job is a little wider in 2026 than it was a few years ago. People still type short queries into a search box, but they also ask full questions, dictate them, and meet your content inside AI Overviews and chat-style assistants that summarize several sources at once. What has not changed is the underlying skill: matching the language of your audience to pages you are actually able to build and maintain. Everything downstream — titles, headings, internal links, content briefs — depends on getting that match right.

Three words get used interchangeably and shouldn’t be. Separating them early makes the rest of the process much easier:

• Query – the exact string a person typed or spoke.
• Keyword – a query you have decided to target with a specific page.
• Topic (or parent topic) – the cluster of related queries that a single page can realistically satisfy.

Is keyword research important?

Yes, keyword research is the best way you could form an idea on what people are typing in search engines when they are looking for content. In order to avoid creating content that is not relevant and not desired , keyword research is a must.

Practically, it is the step that decides where your budget goes. Every article, landing page or product description costs time to write, review, translate and keep current. Research is what stops you from spending that budget on a page that has no demand behind it, or on a query so competitive that nothing short of a two-year link-building program would move it. Google’s own SEO Starter Guide makes the same point in different words: write for the way people actually ask, not for the way you internally describe your product.

A finished round of research should be able to answer five questions in one sitting:

• What does my audience search for, in their words rather than mine?
• Which of those searches can my site realistically compete for this year?
• Which of them are close to money, and which are pure top-of-funnel reading?
• Which pages do I already have that could be improved instead of duplicated?
• What is the order of work for the next quarter?

How to find keyword ideas

Before you collect a single idea, open a spreadsheet with these columns: keyword, source, estimated volume, difficulty, intent, parent topic, business value, existing URL, action. Filling those columns is the whole exercise. Without them you end up with a long, flat list that nobody ever converts into work.

Think about what visitors would search for.

You can start by thinking about revelant words about the topic around which your content is made, then use keyword research tools to find more relevant words for your chosen subject.

There are four simple guidelines to help you find relevant keywords for your page.

1. Think about “seed” keywords.

Seed keywords are the main, most important word you are going to use to help you describe your content and identify your competition. Research tools also use seed keywords to generate a list of other relevant words.

Seeds are not discovered, they are written down from what you already know. Spend twenty minutes listing: the products or services you sell, the categories they sit in, the problems they solve, the names of the tools or brands your customers compare you against, and the words your sales team hears on calls. Ten to twenty seeds is plenty — each one will fan out into hundreds of suggestions.

Worked example. Imagine a small shop selling refurbished espresso machines. The product seeds are espresso machine, refurbished espresso machine, coffee grinder. The problem seeds come from the support inbox: espresso machine leaking, descaling, portafilter pressure. The comparison seeds come from the market: brand names, plus manual vs automatic espresso machine. Three very different families of intent, from one twenty-minute exercise — and each family will need a different type of page.

2. Identify the keywords your competition uses.

First, you need to identify that competition by using the keywords you chose. You can do so by googling those seed keywords and see who ranks on the first page. If you do not find any top ranking sites that are similar to yours, you should try “autosuggest” queries instead.

Once you’ve identified these competitors you can use tools sites like Ahrefs Site explorer to see their top pages report. In this report you will find top keywords for each page.

Two habits make this step far more useful. First, distinguish search competitors from business competitors: the sites beating you in the results are often publishers and marketplaces, not the companies your sales team worries about — and they are the ones whose pages you have to out-do. Second, before you look at anyone else, look at yourself. The Performance report in Google Search Console is the closest thing to first-party keyword data you will get: real impressions, real clicks, real average positions for your own URLs. Sort by impressions, filter to positions 5–20, and you usually find a dozen keywords you are already half-ranking for and never wrote about on purpose.

3. Use keyword research tools.

Keyword research tools are straightforward, you put in a keyword and they generate a list of more keywords that are relevant to yours.

No single source gives you the full picture, so it helps to know what each one is good at and where it lies to you:

Source	Best for	Limitation
Google Search Console, Performance report	Queries you already get impressions for; first-party, no estimates	Only shows what you already rank for
Google autocomplete and People Also Ask	Real phrasing, question variants, long-tail wording	No volume numbers; has to be collected by hand
Competitor top pages reports	Topics with proven demand and a proven format	You inherit their strategy, including their mistakes
Keyword tools (Keywords Explorer, Google Ads Keyword Planner)	Volume, difficulty and clustering at scale	Estimates; Keyword Planner shows broad ranges unless the account is spending
Google Trends	Direction of demand, seasonality, regional differences	A relative index, not absolute search volume
Communities: Reddit, niche forums, your own support inbox	Problems nobody has written about yet	Hard to size; some terms have no measurable volume at all

4. Research and study your chosen subject.

You can start by browsing all matter of forums, Q&A sites and communities. In doing so you might find all sorts of problems and topics your targeted audience might be interested in, and maybe your competitor did not bother to cover.

A quick way to force variety out of a single seed is to run it through modifier families and note what autocomplete offers back: questions (what, why, how, when), comparisons (vs, alternative, instead of), qualifiers (best, cheap, professional, for beginners), commercial terms (price, cost, buy, agency, service) and location terms if you sell locally. Ten minutes of this per seed usually produces more genuinely useful long-tail ideas than an hour of clicking through tool exports.

How to analyze keywords

It is good to have tons of ideas and keywords, but how do you know which are the best?

Use SEO metrics, to narrow the list down and refine your keyword pool. An example of keyword metric is Search volume. You can use search volume to find the average nr. of times a keyword gets searcher per month. But keep in mind that this number is not the number of users who searched but the number of searches. Also it doesn’t tell you how much traffic you will get by ranking and it’s an annual average.

Volume alone is the metric that gets beginners into the most trouble, so it is worth reading a whole row of metrics together rather than sorting a spreadsheet by one column:

Metric	What it actually tells you	The trap
Search volume	Roughly how many searches that query gets per month, usually as a twelve-month average	Counts searches, not people, and not clicks
Clicks / clicks per search	How often a search ends in a click on an organic result	Queries answered on the results page itself — AI Overviews, featured snippets, definitions, unit conversions — can be high-volume and nearly click-free
Keyword difficulty	A vendor’s estimate of how hard the first page looks, mostly from links pointing at the ranking pages	Vendor-specific and not comparable across tools; ignores brand strength, freshness and intent mismatch
Traffic potential	How much traffic the current number-one page gets across all of its keywords	Only meaningful if you can build something comparable
Business value	How close the searcher is to buying from you	Subjective, so score it explicitly instead of deciding by feel
Seasonality	Whether demand is flat or spikes, visible in Google Trends	An annual average hides a three-week peak completely
SERP features	What occupies the page: AI Overview, shopping, local pack, video, forum threads	Sets the realistic ceiling for a classic blue link

Worked example. Suppose your tool reports a few thousand monthly searches for what is espresso and only a couple of hundred for refurbished espresso machine warranty. Sorted by volume the first wins by a mile. Now check the results pages. The first is answered above the fold by an AI Overview and a dictionary-style snippet, and the people clicking are not buyers. The second returns retailer pages from four shops, no AI Overview, and every visitor is a person with a credit card in their hand deciding whom to trust. On volume, one keyword; on revenue per hour of writing, the other. This is the comparison you should run for every candidate before you commit to a brief.

How to target keywords

Each time you’re creating content using keywords, you need to make sure you choose the right page and content for that specific keyword. To do so you can follow two simple guidelines:

• Identify the parent topic

If you have a bunch of keywords on your list and you are wondering how to insert them in your text, should you put them all in one page or should you build a different page for every keyword? Well, it depends. It all depends on how google sees these words. Are they part of the same topic or are they individual topic keywords? You can determine this by searching your words and see how they rank. If two of your similar keywords are ranked on the same page after your search , that means they are similar topics, and you can insert them in the same page. You can also use Keyword explorer to identify a parent topic for every keyword. Write your words in Keyword explore and it will list the keywords which bring the most traffic to the site, and those are your “Parent Topic” keywords.

The manual version of this check takes about a minute per pair: search keyword A, search keyword B, and compare the top ten results. If most of the same URLs appear in both, one page can serve both. If the two result sets barely overlap, they are separate topics and deserve separate pages. Getting this wrong in the direction of “a page per keyword” is how sites end up with five thin articles about the same thing, all competing with each other and none of them ranking — the problem usually called keyword cannibalization. If you suspect you already have it, filter the Search Console Performance report by that query and switch to the Pages tab: if the impressions for one query are split across several of your URLs, consolidate them into the strongest one and redirect the rest.

• Identify search intent

Let’s say you have this keyword in mind, “best football balls”, and you want to sell your ball collection, that won’t work. When people search for “best football balls” they usually want to see a list with best balls for football not a site that sells balls, but of course that is not always necessarily true. In any case google knows search intent best, so if you want to create content suitable for users and to match your desired keywords, then you should, again, google those keywords and see what type of pages rank on the first page of the results. If your words give you blogs or lists as results, then you should either change your keywords a little bit, or adapt and build the same type of content. Take your time to familiarize yourself with this concept because it is very important to align with the search intent if you want your page to rank high in the results page.

Most queries fall into one of four buckets, and each bucket wants a different kind of page:

Intent	Typical query shape	What usually ranks	What to build
Informational	what is, how to, why does, guide	Tutorials, definitions, explainers	An article or how-to, linking on to your commercial pages
Commercial investigation	best, top, review, X vs Y, alternatives	Listicles and comparisons	A roundup or comparison with a clear method
Transactional	buy, price, cost, quote, agency, near me	Product, category and service pages	A landing or service page, not a blog post
Navigational	a brand name, login, pricing page of a specific tool	That brand’s own pages	Nothing, unless the brand is yours

When you have the bucket, read the top results once more for three things: the content type (article, category page, tool, video), the format (step-by-step, listicle, comparison, opinion) and the angle (for beginners, for 2026, cheapest, fastest). Matching all three is what people usually mean when they say a page “matches intent”. It also happens to be exactly what Google describes in its guidance on creating helpful, reliable, people-first content.

How to prioritize keywords

While you go through the steps above, finding your keywords, analyzing and grouping them, also think about prioritizing them. You should pick the ones that are most suitable for you after all. Ask yourself, do you already have the content for those topics, or what will it take to create content for the said topic, how much traffic will it bring and how tough is the competition.

Turn those questions into a score so the decision survives contact with other people’s opinions. Rate every parent topic from 0 to 3 on four axes — demand, business value, difficulty relative to your site, and effort to produce — then rank by (demand + business value) − (difficulty + effort). It is crude, and that is the point: it is fast, repeatable, and it makes disagreements about a topic explicit instead of political.

Parent topic	Demand	Business value	Difficulty	Effort	Verdict
Broad head term, e.g. “espresso machine”	High	Low	High	High	Park it; revisit when the domain is stronger
Comparison term, e.g. “manual vs automatic espresso machine”	Medium	Medium	Medium	Medium	Schedule for next quarter
Bottom-funnel term, e.g. “refurbished espresso machine warranty”	Low	High	Low	Low	Write this week
Support question already ranking on page two	Low	Medium	Low	Very low	Update the existing page, don’t write a new one

Notice the pattern: the fastest wins are usually improvements to pages you already own. Refreshing a page that sits in positions 5–15 is cheaper than commissioning a new one, and Search Console tells you exactly which pages those are.

From keyword list to publishing plan

Research that stops at a spreadsheet is research that never earns anything. The handover step is short:

1. Group the list by parent topic, one row per future page.
2. Map each group to a URL — an existing one to update, or a new slug to create. Never leave a group unmapped.
3. For each new page write a one-page brief: the target parent topic, the variants to cover, the intent bucket, the format, the questions from People Also Ask you will answer, and the internal links in and out.
4. Set the order of publication from the priority score, not from whichever topic is most fun to write.
5. Record the starting position and impressions from Search Console on the day you publish, so that in ninety days you can tell whether the page worked.
6. Re-run the whole exercise quarterly, and check Search Console monthly for queries you never planned for.

Common keyword research mistakes

• Sorting by volume and stopping there. The biggest number in the sheet is usually the hardest and the least commercial.
• One page per keyword. Variants of the same topic belong together; splitting them starves every version.
• Ignoring the results page. If the top ten are all category pages, a blog post will not break in, no matter how good it is.
• Chasing terms your site cannot support. Difficulty is relative to your own authority, not an absolute.
• Researching once. Demand moves, competitors publish, and search interfaces change; a list from two years ago is a historical document.
• Forgetting the other languages. If you operate in more than one market, translate the intent, not the keyword — the literal translation of a term is frequently not what people in that market search for.

Keyword research FAQ

How many keywords should one page target?

One parent topic, and as many variants of it as the page can honestly answer. A well-built guide typically picks up dozens or hundreds of long-tail queries it was never explicitly optimized for, simply because it covers the subject properly.

Is keyword density still a ranking factor?

No, and repeating a phrase to hit a percentage makes writing worse without making rankings better. Aim to cover the subtopics a reader expects, use the natural wording in your title, headings and first paragraph, and then write like a human.

Should I target keywords with zero search volume?

Often, yes. Tools round low numbers down to zero, brand-new terms have no history, and very specific bottom-of-funnel questions can be worth more than a thousand casual visits. Judge them on business value, not on the number in the volume column.

Do AI Overviews and AI assistants make keyword research pointless?

They change what a ranking is worth, not whether research is needed. Generated answers are assembled from pages that already cover the topic clearly and are easy to parse, so the work shifts toward topics where a click still has a purpose — comparisons, buying decisions, local services, anything requiring trust — and toward writing pages that answer a question outright rather than burying the answer.

Do Core Web Vitals affect which keywords I can win?

Indirectly. Page experience is not a substitute for relevance, but on competitive terms where several pages answer the query equally well, a slow page is an easy one to drop. Note that the metrics themselves moved on: First Input Delay was replaced by Interaction to Next Paint as a Core Web Vital in March 2024, so old checklists that still mention FID are out of date.

How often should I redo the research?

A full pass once or twice a year is usually enough for a small site, with a monthly fifteen-minute look at Search Console to catch new queries, seasonal shifts and pages slipping off page one.

Tool for keyword research

If you read through all the steps above, you are one step closer to create content for the first page of google, but of course, practice makes perfect, so don’t expect results without work. 

To wrap this up I have some bonus goodies for you, some useful links with research tools, useful for researching keywords.

• Google trends.
• Google search console.
• Ahrefs webmaster tools.
• Keyword generator.
• Keywords explorer.
• Keyword rank checker.
• Keyword difficulty checker.

Two housekeeping notes, because tool names age badly. The keyword tool inside Google Ads is the Keyword Planner — it was the AdWords Keyword Planner until Google retired the AdWords name, and it still groups volumes into wide ranges for accounts that are not spending. On the measurement side, Universal Analytics stopped processing data in 2023, so the landing-page and conversion reports that tell you which keyword clusters actually pay are now in GA4. Feature sets and free tiers change often, so check what a tool offers today rather than trusting any roundup, including this one.

If you would like a second opinion on a keyword list before you commission the content — or help building the authority that makes competitive terms reachable in the first place — that is the part of the job we do every day at web-push.com.