How to Do Keyword Research: A Practical Guide for Growing Businesses

Most businesses that ask how to do keyword research already have a spreadsheet full of guesses. The list usually has fifty terms on it. Half are too broad to rank for, and half are too obscure to matter. The real skill isn't generating keyword ideas. It's picking the ones that bring in people ready to buy, not just visitors who bounce.

Getting this wrong is expensive in a way that doesn't show up until months later. A team spends six weeks writing content around the wrong terms and sees no traffic. They conclude SEO doesn't work for their industry. In most cases, the industry was fine. The keyword list was the problem.

This guide walks through the process we run for clients when we do keyword research. It covers how to find the right terms, sort them by what they'll actually do for revenue, and turn the list into content that ranks.

How to do keyword research: the core process

Keyword research exists to answer one question: what is your buyer typing into Google before they find you? Or before they find a competitor instead? Get that question right and everything downstream (content, ad targeting, site structure) gets easier. Get it wrong and you're producing content nobody was searching for.

A lot of businesses skip this step and write about what they think is interesting. That approach produces blog posts with zero search volume and zero traffic. The businesses that get this right start from demand data, not internal opinion, before a single page gets written.

Three things a solid keyword list should tell you:

• What your buyer calls the problem. Not your internal jargon for it, their actual words. A dentist's patient searches "tooth pain when I bite down," not "occlusal trauma."
• How close they are to buying. Someone searching "what is SEO" is months from a purchase. Someone searching "SEO agency pricing" might convert this week.
• Where competitors are weak. If three competitors rank for a term with thin, outdated content, that's an opening.

Start with search intent, not search volume

Search intent tells you why someone typed a query. It matters more than volume, because a high-volume term with the wrong intent will never convert, no matter how well it ranks. Match the wrong content to the wrong intent and you'll rank for nothing that pays the bills.

Google groups intent into four rough categories, and each one calls for different content:

• Informational. The searcher wants an answer, not a purchase. "How long does SEO take to work" is informational. Blog posts and guides fit here.
• Navigational. The searcher already knows where they're going, they're just using a search box to get there. "Aim Growth login" is navigational.
• Commercial. The searcher is comparing options before deciding. "Best SEO agency for small business" is commercial.
• Transactional. The searcher is ready to act. "Hire SEO consultant" or "SEO pricing" is transactional.

A common mistake: writing an informational blog post, then bolting on a hard sales pitch at the end. The reader came for an answer, not a quote. Save the direct pitch for pages built around commercial and transactional intent. Let informational content do its job instead, which is earning trust before the reader is ready to buy.

This is where a lot of businesses do keyword research backward. They find a high-volume term, write toward it, and only afterward notice the intent doesn't match anything they sell. Checking intent first, before a single word gets written, saves the rewrite later.

The keyword research tools worth paying for

You need three kinds of data to do keyword research properly. That means search volume, ranking difficulty, and the real questions people ask around your topic. No single free tool gives you all three reliably. Most serious research combines two or three sources, then cross-checks the results against what the site already ranks for.

• A paid keyword database (Ahrefs, Semrush, or similar) for volume and difficulty scores. These aren't exact, but they're consistent enough to rank opportunities against each other.
• Google Search Console, if the site already has traffic. It shows queries you're already ranking for on page two or three. Those are often the fastest wins, because you're already halfway there.
• The search results themselves. "People also ask" boxes and related searches surface real phrasing. So do forum threads on Reddit and industry-specific communities, which keyword tools sometimes miss entirely.

Running this process in-house works fine if someone owns it consistently. It falls apart when it's the fourth priority on someone's list and gets revisited once a year. If that's the pattern at your business, our SEO services cover this end to end. That includes the research and the work of prioritizing which terms are worth building content around first.

Finding long-tail keywords your competitors ignore

Long-tail keywords are longer, more specific search phrases with lower individual volume but a much higher rate of converting. Someone typing five or six specific words already knows exactly what they want. A business chasing only high-volume head terms fights for the most competitive real estate on the page. Meanwhile, dozens of easier wins sit nearby, ignored.

Three places long-tail terms hide that most keyword tools won't surface directly:

• Your own sales calls. The objections and questions your sales team hears on repeat are search queries in disguise. If five prospects this month asked "how much does a website redesign actually cost," that's a title. Not just a talking point.
• Support tickets and reviews. Customers describe problems in their own words here. That's exactly the phrasing search engines reward when it matches what people type.
• Question-based tools. AnswerThePublic and the "People also ask" box both cluster real questions around a seed term. Plenty of those questions have almost no competing content written for them yet.

A page targeting "website redesign checklist" will rarely out-rank a page targeting "how much does a website redesign cost for a 20-page site." The second term gets a fraction of the searches. But it matches a much narrower, further-along buyer.

This is also where a lot of the volume in a properly built keyword list actually lives. A single head term might carry ten times the search volume of any one long-tail variant. But twenty long-tail variants around that same topic often add up to more combined traffic, at a fraction of the competition.

Where AI search changes keyword research

Search no longer starts and ends on Google. ChatGPT, Perplexity, and Google's own AI Overviews now answer questions directly. They pull from pages that cover a topic clearly, not pages stuffed with one exact-match phrase. Keyword research now has to account for how AI systems parse and cite content, not only how Google's algorithm ranks it.

That shifts the research process in two ways. First, topic clusters matter more than individual keyword variants. An AI engine summarizing "how to do keyword research" will pull from a page that answers the surrounding questions too: tools, intent, long-tail strategy. Not a page that repeats the same five-word phrase forty times. Second, the phrasing of the query itself is changing. People type "what's the best way to find keywords for a new website" into a chat window. They'd never type that into a Google search bar, and that longer, conversational phrasing needs its own research pass.

We cover the mechanics of ranking in these systems in our guide on how to rank in ChatGPT search. It's a big enough shift that our AEO and generative engine optimisation work now runs alongside traditional keyword research for most clients, rather than as a separate project.

Practically, that means every keyword research project should now produce two outputs, not one. One is a list of terms to target on the page. The other is a map of the questions and subtopics an AI engine needs answered to cite that page as a source. The two overlap heavily, but they aren't identical. Skipping the second one means writing content that ranks on Google but never gets pulled into an AI-generated answer.

Turning keyword research into a working content plan

A keyword list is only useful once it's sorted and assigned. Score each term on three factors: search volume, ranking difficulty, and how close the intent is to a purchase decision. Then build content in that order. Start with the terms that combine reasonable volume, low-to-moderate difficulty, and buying intent close enough to matter.

Prioritizing correctly:

1. Rank every keyword on a simple 1-5 scale for volume, difficulty, and intent proximity to purchase.
2. Start with commercial and transactional terms your site doesn't yet cover. These convert fastest once they rank.
3. Layer in informational terms that support the commercial ones, feeding internal links back to service pages.
4. Revisit the list quarterly. Search behavior shifts, especially with AI search adoption changing how people phrase questions.

Set expectations on timing before committing resources. Ranking is not instant, even for well-researched terms. We've laid out realistic timelines in our post on how long does SEO take to work, worth reading before a content calendar gets built around unrealistic dates.

Where to start this quarter

Pick ten keywords, not a hundred. Ten well-researched, intent-matched terms with a clear content plan will outperform a hundred-term spreadsheet nobody ever finishes acting on. That's the whole point of learning how to do keyword research properly: fewer terms, chosen better, beat a long list nobody prioritizes. Start with the terms closest to a buying decision and build outward from there. Revisit the list every quarter as search behavior shifts.

If the research and prioritization is the part that keeps stalling, get in touch. We'll walk through where the fastest wins are likely sitting in your market.