Keyword Research Fundamentals: Finding the Terms Your Customers Actually Search

Most keyword research still starts backwards. A business decides what it wants to say, checks whether the phrase gets any search volume, and calls that “research.” Real keyword research starts from the other direction: what is a potential customer actually typing into Google before they’ve ever heard of your company, and does your content answer it better than what already ranks?

This matters more for a services business than for almost anything else, because the words a customer uses before they buy are rarely the words the business uses to describe itself internally. A company sells “server administration.” A customer searches “why is my website so slow.” A company sells “on-page SEO.” A customer searches “why did my traffic drop after I redesigned my site.” Keyword research is the process of closing that gap.

What Keyword Research Is Actually For

Keyword research gets treated as a box-ticking exercise before writing: pick a term, hit a target density, move on. That version produces content that reads like it was written for a spreadsheet. The useful version does three things before a single sentence gets written:

• It tells you whether anyone is searching for this at all, and how many.
• It tells you what they mean when they search it — the intent behind the words.
• It tells you what already ranks, so you know what you’re competing against and whether you can realistically win.

Skip any one of those three and the research is decorative. A term with high volume and the wrong intent will get clicks that bounce immediately. A term with the right intent and zero competition analysis will get you outranked by a page ten times more thorough than yours. A term that scores well on both but has no real search volume will never bring anyone to the page.

Search Volume Is a Direction, Not a Target

Volume numbers from any tool — Google’s own Keyword Planner included — are modeled estimates, not counts. They’re useful for comparing two phrases against each other, and nearly useless as an absolute number to plan a business around. A term showing “10–100 searches a month” might get 15 or might get 95; the tool genuinely can’t tell you which, and neither can a competitor’s marketing blog. Treat volume as a way to rank options, not a number to defend in a meeting.

Search Intent Decides Whether You Can Win at All

Every query carries an implicit job the searcher wants done. Broadly, that job falls into one of four buckets:

• Informational — “what is server hardening”, “how does live chat work”
• Navigational — “postrss login”, “talkmio pricing”
• Commercial investigation — “best live chat tool for small business”, “postrss vs alternatives”
• Transactional — “hire seo agency”, “buy live chat widget”

Before writing anything, check what Google is actually showing for the term. If a query returns ten in-depth explainer articles and your page is a service landing page, you’re fighting the algorithm’s own read of what searchers want. If it returns product comparison pages and yours is a glossary definition, same problem in reverse. The fastest way to check intent isn’t a tool — it’s opening the query in an incognito window and looking at what’s already on page one.

A Practical Research Process

1. Start From Questions, Not Words

Instead of guessing a keyword, write down the actual questions your sales and support conversations keep answering. Every recurring question a prospect asks before signing is a query someone else is typing into Google right now. A company running IT maintenance contracts hears “what happens if my server goes down at night” far more often than it hears anyone say “server administration services” out loud.

2. Expand With Free Tools Before Paid Ones

Google’s autocomplete and the “People also ask” and “Related searches” panels on the results page are free, current, and reflect real query patterns — not modeled averages. Google Trends adds a time dimension: whether a topic is rising, seasonal, or flat helps decide whether to invest in an evergreen piece or a timely one. Paid tools add search volume estimates and difficulty scores on top of this, which are helpful for prioritization but not required to get started.

3. Group Queries by Intent, Not by Similarity

A common mistake is building a list of forty keyword variants of the same phrase and trying to write one page that targets all of them. Google doesn’t reward keyword density anymore; it rewards a page that thoroughly answers one intent. Group your list into clusters that share intent — one cluster becomes one page, not forty thin pages or one bloated page trying to rank for everything.

4. Check What You’re Actually Competing Against

Search the term and look honestly at the top five results. Are they from established domains with years of backlinks, or from pages you could realistically outwrite? Is the format a listicle, a tool, a long guide, a forum thread? Matching the format Google already favors for that query is often more important than matching the exact keyword phrase.

Keyword Research for a Services Business vs. a Product Business

The mechanics are the same but the priorities shift depending on what’s being sold.

Signs Your Keyword List Needs Rework

Every Term Has the Company Name Already Baked Into the Thinking

If every keyword idea starts from a service page title rather than a customer problem, the list will be full of terms nobody searches. Rebuild it from support tickets, sales call notes, and the actual phrasing customers use — not the phrasing used internally.

The List Ignores What’s Already Ranking on the Site

Before targeting a new term, check whether an existing page on the site is already ranking for something close to it. Publishing a second page on a near-identical topic doesn’t add a second chance to rank — it usually splits authority between two pages and can quietly cost both of them position. A short internal search across the site’s existing content, both by title and by topic, should happen before a new brief is written, not after the draft is finished.

Nobody Has Checked Difficulty Against the Site’s Actual Authority

A five-year-old site with a shelf of backlinks can chase competitive terms that a six-month-old blog cannot. Realistic keyword research includes an honest look at whether the current site has any chance of outranking page one, and if not, whether a narrower, more specific version of the term is winnable instead.

Turning Research Into a Brief

A keyword list is not a content plan until it’s translated into a brief that says: this is the primary term, this is the intent it serves, these are the sub-questions the page needs to answer to satisfy that intent completely, and this is the format that already works for this query on page one. Writers who get a bare keyword instead of this context tend to produce pages that mention the term without actually resolving the reader’s question — which is precisely the kind of page that ranks for a week and then slides.

Once a page is live, the work isn’t finished. A structured on-page checklist makes sure the research actually gets reflected in titles, headings and internal links, and pairing that with the practice of reading the resulting search performance report closes the loop — showing which of the original keyword guesses were right and which need revisiting next quarter.

FAQ

How many keywords should one page target?

One primary term and a handful of closely related phrases that share the same intent. If two target terms would require answering different questions in different formats, they belong on two separate pages, not one page trying to do both jobs.

Do exact-match keywords still need to appear word-for-word in the text?

No. Modern search systems match meaning and topic coverage, not exact strings. Writing naturally about the topic and its sub-questions matters more than repeating a phrase verbatim.

Is low search volume a reason to skip a topic?

Not automatically. A handful of highly qualified visitors searching a specific, high-intent phrase can be worth more than a much larger volume of visitors searching a vague, informational one. Volume is one input, not the only one.

How often should keyword research be redone for an existing page?

Revisit it when performance data — impressions, clicks, average position — suggests the page is being shown for different queries than originally intended, or roughly once a year for pages covering a topic that evolves, such as anything tied to a specific platform or advertising product.

What’s the fastest way to check search intent without a paid tool?

Open an incognito browser window and search the exact phrase. What Google already shows on page one — guides, tools, comparisons, local listings — is the clearest available signal of what it believes searchers want.

Should keyword research include competitor websites?

Yes, but as a check on difficulty and content gaps, not as a template to copy. Seeing what competitors rank for reveals both terms worth targeting and terms already saturated by better-resourced sites.

Can keyword research be done without any paid software?

Yes. Google Search Console (for a site with existing traffic), autocomplete, “People also ask,” Google Trends, and a careful read of the current top-ranking pages cover the fundamentals. Paid tools speed up prioritization at scale but aren’t required to start.

The Bottom Line

Keyword research that works starts with the customer’s actual words, not the company’s internal vocabulary, and it ends with a brief a writer can act on — not just a list of phrases to sprinkle into a draft. Volume tells you scale, intent tells you whether you can win, and a look at what’s already ranking tells you what “winning” will actually require. Skipping any one of those turns research into decoration.