Keyword Research Basics: Pick Keywords You Can Actually Win

Keyword research basics come down to four moves: brainstorm the words your customers type, check each term's search volume, difficulty and intent, keep the ones you can realistically win, and assign every winner its own page. Do that before you write anything, and every page you publish starts with demand already proven.

The stakes are simple. According to an Ahrefs study of over one billion pages, 90.63% of pages on the internet get no organic traffic from Google. Most of those pages were not unlucky. They were written about topics nobody searches for, or aimed at terms their sites could never rank for. Research is how you avoid joining them.

This guide is for site owners, marketers and founders doing their own SEO, on any kind of site, in any market. Ahrefs estimates you can learn most of the discipline in about 20 minutes. Working through your own first list takes an afternoon, and it shapes everything you publish afterward.

One thing worth saying plainly before the method: keyword research basics are a skill to learn, not a product to buy. If you typed the phrase into a search engine, you wanted a clear explanation you could act on, and that is what this page is. Nothing below requires a purchase; every step works with free tools.

What keyword research is and why it decides your traffic

Keyword research is the process of finding, analyzing and selecting the words and phrases people type into search engines, so you can build content that matches real demand. The keyword is the gateway: it connects a searcher's question to your page. Pick the right terms and you earn a steady stream of visitors who were already looking for what you offer. Pick wrong, or skip the step, and you write for an audience of zero.

You need it at three moments: when you enter a new niche, when you plan new content, and when you improve pages that already exist. In every case the job is the same. Learn the language your potential visitors actually use, then meet them in it.

The payoff is commercial, not academic. One SEO consultant writing for HubSpot reports growing client blog revenue by 254% year over year with an intent-led keyword process. The reasoning behind that kind of result, and how research turns into money rather than mere traffic, is laid out in why keyword research is essential for SEO results that pay. And if you are brand new to SEO as a whole, the complete beginner's guide to finding keywords for SEO covers the wider context this article builds on.

The three numbers on every keyword

Every keyword tool reports the same core metrics. Read all three before you commit to a term, because each one can veto the other two.

Metric	What it tells you	Working benchmark	Where beginners go wrong
Search volume	Average monthly searches for the term	100-1,000 monthly searches is the sweet spot for most new sites, per HubSpot's guide	Chasing 10,000+ head terms with no site authority
Keyword difficulty	How hard ranking will be, scored 0-100 from the strength of pages already ranking	Under 40 for new sites, again per HubSpot's benchmarks	Treating a high score as a hard no instead of a longer timeline
CPC (cost per click)	What advertisers pay for a click on that term	High CPC signals commercial value even if you never run ads	Ignoring it because you only do organic

Search volume: how much demand exists

Volume is average monthly searches. Higher is not automatically better. HubSpot's benchmarks are a useful ladder: 10-100 searches a month is low but often highly specific and convertible, 100-1,000 is the sweet spot for most businesses starting out, 1,000-10,000 is competitive but worthwhile for established sites, and 10,000+ is usually dominated by major brands. A term with 500 searches that you can actually rank for beats a term with 10,000 that you cannot.

Keyword difficulty: how hard the fight is

Difficulty scores run 0-100, calculated mostly from the authority of the sites currently ranking. HubSpot's bands: 0-30 means new sites can compete, 31-60 needs some existing authority and quality content, 61-80 requires strong domain authority and backlinks, and 81-100 belongs to major brands. Real terms sit everywhere on this scale. In a published walkthrough, Iowa State's web team found "statistics minor" drawing 590 monthly US searches at a 50% difficulty rating, a mid-tier fight a university site could take on but a brand-new blog should not.

CPC: what the click is worth

Cost per click is an advertising number with an organic lesson. HubSpot's guide puts it plainly: if businesses pay $20 per click for a keyword, that keyword drives revenue. A term with high CPC and high advertiser competition is a money keyword, and ranking organically for it means capturing clicks others pay for daily. You can pull volume, difficulty and intent for any term with a free keyword checker before deciding whether it earns a page.

How to find keywords: five sources that fill a list

Your goal at this stage is quantity of relevant ideas. Filtering comes later. These five sources, worked in order, will give you more candidates than you can use.

Start with seed keywords from your own business

Put yourself in your customer's position and write down what they would type. A coffee equipment seller's seeds might be "french press", "espresso machine" and "pour over kettle". Do not overthink the list; seeds exist to be expanded. A free expander turns each seed into dozens of ideas: Ahrefs' free keyword generator, for example, returns 20 related terms and 20 questions per seed, 40 ideas from one word. A free keyword suggestions tool does the same job with a focus on long-tail phrasings. AI chatbots brainstorm seeds quickly too, with one caveat: they return no real volume or difficulty data, so every idea still needs checking in a tool.

Read the results page before you trust a term

Search each promising keyword and study what Google returns. Practitioners call this search listening. The results page tells you what format wins (videos, product pages, guides), who you would be competing against, and what related questions people ask. The People Also Ask box alone often adds four or five new candidates per search. It also reveals your audience's thought process: someone searching "is a statistics minor useful" is weighing a decision, not browsing, and the page that ranks should speak to that.

Mine your own Search Console data

If your site has history, Google Search Console lists the queries you already rank for and get clicks from. These are your cheapest wins, because improving a page from position 15 to position 5 is far easier than ranking from nothing. The catch: Search Console shows no volume or difficulty, and Ahrefs notes it caps the query list at your top 1,000 keywords, so cross-check anything interesting in a proper tool.

Borrow from competitors

Search your seed terms and note which sites keep appearing, then make sure they are true peers. If you sell coffee equipment, a rival equipment store is a competitor; a cafe chain is not. Competitive tools can then show which keywords bring those sites their traffic, and a content gap check reveals terms they rank for that you do not. Different situations call for different sweeps, and the eight types of keyword research and when to use each maps which one fits the moment.

Listen where customers actually talk

Forums, subreddits and your own customer conversations surface phrasings no tool suggests. Ahrefs gives a telling example: a Reddit thread about brewing ratios pointed to the query "aeropress coffee to water ratio", which shows only 150 searches a month, yet the thread's 42 upvotes proved people genuinely value the answer. Low-volume terms backed by visible enthusiasm often convert far above their size. When you are ready to turn a raw list into rankings, how to conduct keyword research that actually ranks walks the full workflow end to end.

Long-tail keywords do the quiet work

A long-tail keyword typically runs three or more words, carries lower search volume and lower competition, and signals a specific need. Moz's beginner guide estimates that head terms make up only a fraction of all searches, with the remaining 75% sitting in the "chunky middle" and long tail. That is where most realistic opportunities live.

Specificity is the reason they convert. A search for "running shoes" is a browse with ambiguous intent; Google cannot tell whether the person wants reviews, a store, or a definition. A search naming the exact model, size and a price cap is a purchase in progress. The same logic holds in every niche: "hiking" is unwinnable and vague, while "how many calories does hiking burn" is a precise question one good page can own.

One warning: long-tail does not mean one blog post per phrasing. There are near-infinite variations of any question, and modern search engines group them. Write one thorough page per subtopic and it will collect rankings for dozens of long-tail phrasings you never typed. Chase each variation with its own thin page and you split your effort into fragments that all lose.

Match the intent before you write the page

Search intent is the reason behind the query, and it decides what kind of page can rank. Google's Quality Rater Guidelines describe intents as know, do, website and visit-in-person, which map closely to the four types most SEO tools report. Get intent wrong and even perfect optimization fails, because your page answers a question nobody asked.

Intent type	What the searcher wants	Example query	Page that matches
Informational	To learn or research something	"how does keyword difficulty work"	Blog post, guide or FAQ page
Navigational	To reach a specific site they already know	A brand name plus "login"	That brand's own page; rarely worth targeting
Commercial	To compare options before choosing	"best keyword research tools"	Comparison page, category roundup or review
Transactional	To act or buy now	"buy ergonomic desk"	Product, service or pricing page

The fastest intent check is the results page itself. If the top results for your term are all comparison listicles, Google has decided the intent is commercial, and your product page will struggle there no matter how well built it is. Match the dominant format first, then out-write it. Local businesses get a fifth pattern: visit-in-person searches such as "bakery near me" return map results, and only location-ready pages compete in them.

Build a keyword map: one page, one keyword

A keyword map assigns each priority keyword to exactly one URL, along with the data you used to choose it. This is the deliverable of all keyword research, the sheet that turns a list into a publishing plan. Iowa State's web team published a clear example of the format:

Keyword	Monthly searches	Difficulty	Intent	Target page
statistics major	1,000	55	Informational	Undergraduate major page
statistics degree	590	50	Informational and commercial	Undergraduate major page
statistics minor	480	28	Informational	Undergraduate minor page

Notice two things. Close variants like "statistics major" and "statistics degree" share one page, because they mean the same thing to a searcher. And no keyword appears on two rows with two different pages. That second rule prevents keyword cannibalization, where your own pages compete against each other in results, split impressions and clicks between them, and often both rank worse than either would alone. When you catch cannibalization, merge the pages into one stronger piece.

Two habits make the map stronger. First, cluster before you assign: grouping a raw list by topic with a free keyword clustering tool shows which phrasings belong on the same page and which genuinely need their own. Second, note seasonality. Moz gives the example of "christmas box" searches spiking from October through December in the UK; a keyword with a season needs its page published months ahead. Once the map exists, the work shifts from research to execution, and these seven steps from research to rank cover that second half.

Remember also that whole websites do not rank; pages do. Moz makes this point emphatically. Outside the biggest brands, most sites earn more traffic on inner pages than the homepage, which is exactly why each page needs its own uniquely valuable keyword.

Free keyword research tools and what they hold back

You can complete a genuine first pass with free tools alone. You just need to know what each one withholds, because the gaps are where beginners get misled.

Tool	What it gives you	The catch
Google Keyword Planner	Hundreds of keyword ideas straight from Google, plus advertising data	Volumes show as broad ranges like 1K-10K unless you spend on ads, and its competition column measures ad bidding, not organic difficulty
Google Trends	Direction and seasonality of interest over time, by region	Covers big, broad topics only; it will not evaluate specific keyword ideas
Answer the Public	Question phrasings pulled from autocomplete data	The free tier allows three searches a day
Google Search Console	The real queries your site already ranks and gets clicks for	No volume, no difficulty, and the query list is capped
AI chatbots	Fast, unlimited brainstorming of seeds and angles	No real metrics at all; every suggestion needs verification in a data tool

The pattern across all five: free tools are strong at generating ideas and weak at judging them. That judgment layer, reliable volume, difficulty and intent in one place, is what paid tools sell. When the manual process stops scaling, this comparison of keyword research tools that fit an automated SEO workflow shows what to look for, and a keyword engine that gives every keyword its own page closes the gap between finding terms and actually publishing against them.

Common keyword research mistakes that waste months

Most failed SEO projects fail here, not in the writing. Knowing keyword research basics and applying them are different things, and these are the traps that cost the most time.

1. Stuffing the keyword into the page. Repeating a term as many times as possible stopped working years ago. Google's spam policies target exactly this pattern, and pages that read unnaturally lose to pages that read well.
2. Targeting one keyword with multiple pages. Cannibalization splits your impressions and clicks across competing URLs. One keyword, one page, always.
3. Choosing by volume alone. A 10,000-search term at difficulty 85 will send a new site nothing for years. The balance of volume against difficulty is the whole skill.
4. Ignoring intent. A service page aimed at an informational query, or a blog post aimed at a buying query, mismatches what Google has already decided the searcher wants.
5. Skipping the results-page check. The live results are Google showing you its current answer. Publishing without looking means competing blind.
6. Researching once and never returning. Rankings, search behavior and competitors all move. Monitor positions and revisit the map on a regular cycle, because a plan you never re-check cannot teach you what worked.

How AI search changes keyword research basics

Search behavior is genuinely shifting. HubSpot's research into how people search found that 31% use social media to search and 12% use AI chatbots, while noting the caveat that 88% of people still use Google. So the ground is moving, but slowly, and Google remains where most demand lives. You can now rank at the top of a results page and still lose clicks to an AI Overview that answers the question directly on the page, a shift HubSpot's guide calls out explicitly.

This is the latest step in a long trend, not a break from it. Google has spent over a decade moving from matching words to understanding meaning: Panda punished thin content, Penguin punished unnatural keyword use, Hummingbird improved semantic search, and in 2015 Google announced that RankBrain, its AI-based component, had become the third most important ranking signal. Each step rewarded the same behavior: cover the searcher's actual need, not the literal string.

AI search accelerates that. Queries are becoming conversational, so a searcher who once typed "best hiking boots" may now ask a full sentence naming the trip, the season and the terrain. The practical consequence for your research: think in topics and questions, not only exact phrases. A page that covers a subtopic thoroughly can rank, and be cited by AI systems, for phrasings that appear nowhere in its text. Google's own guidance, published on Google Search Central, stays consistent through all of it: make sure content can be crawled and indexed, follow SEO best practices, and create helpful, reliable, people-first content.

None of this makes the manual skills obsolete; it raises the cost of doing them slowly. If a full sweep by hand takes days you do not have, automated keyword research that takes a seed to a site plan in minutes shows what the compressed version of this exact process looks like.

FAQs

What is keyword research in simple terms?

Keyword research is finding out which words and phrases people type into search engines, measuring how many search each one and how hard ranking would be, then choosing the terms your site will target. It ensures you build pages for demand that provably exists.

How long does keyword research take for a beginner?

Learning the method is fast; Ahrefs estimates most of it takes about 20 minutes to pick up. Building your first real list, checking volume, difficulty and intent, and mapping terms to pages typically takes an afternoon. After that, it becomes a recurring habit rather than a one-time project.

How many keywords should one page target?

One primary keyword per page, plus its close variants and related long-tail phrasings, which a thorough page collects naturally. Never point two of your own pages at the same primary term; they will compete with each other and usually both rank worse.

What keyword difficulty should a new website target?

HubSpot's benchmarks suggest staying under 40, and the 0-30 band is where brand-new sites can genuinely compete. Pair that with terms in the 100-1,000 monthly search range and you get keywords that are both winnable and worth winning.

Is keyword research still worth doing now that AI answers questions?

Yes. HubSpot's research found 88% of people still use Google, and AI systems build their answers from content that covers topics clearly and specifically. The work shifts toward topics, questions and intent rather than exact-match strings, but the discipline of validating demand before publishing matters as much as ever.

Map your first twenty keywords

Pick one seed term from your business today, expand it to twenty candidates, and score each for volume, difficulty and intent before you write a single page. That one sheet turns keyword research basics from reading into a publishing plan you can execute this month. If you would rather see the finished product before you start, talk to the team and open a sample SEO report built from a real keyword map to see the difficulty calls, the page assignments and the plan side by side.

Resources / References

1. an Ahrefs study of over one billion pages
2. Moz's beginner guide
3. Quality Rater Guidelines
4. spam policies
5. Google Search Central

Sources

• The Basics of Keyword Research | CALS/LAS Web Team
• How to Do Keyword Research for SEO (Start to Finish)
• [Keyword Research for SEO: The Beginner's Guide 2026)
• What a decade in SEO taught me about keyword research ...
• [Keyword Research Beginner's Guide to SEO)