Start before you open a keyword tool

The best keyword research I’ve done for clients over 3,500+ projects rarely started in a tool. It started with a list of real questions and phrases pulled from actual customer conversations, support tickets, sales calls, reviews left on the business, even competitor reviews. This matters because customers describe their problems in their own words, not in your industry’s internal terminology, and a keyword tool can’t tell you which words those are. It can only expand a list you’ve already started.

Spend twenty minutes before touching any software. Write down every question a customer has actually asked you, in their words. Write down what you’d type into Google if you had the problem your business solves and no idea what it was called. This raw list, unpolished, is often more useful than anything a tool produces on its own.

Understand search intent before you collect a single keyword

Every keyword sits behind one of four basic intents, and getting this wrong is the single biggest reason keyword research leads to the wrong content later.

Informational intent means the searcher wants to learn something: “how does compound interest work,” “what causes lower back pain.” Navigational intent means they’re looking for a specific site or brand: “Nike running shoes,” “Ahrefs login.” Commercial investigation means they’re comparing options before deciding: “best budget laptops 2026,” “Ahrefs vs Semrush.” Transactional intent means they’re ready to act: “buy running shoes size 10,” “book dentist appointment near me.”

Before you add a keyword to your list, glance at what’s currently ranking for it in Google. If the results are all product pages and you were planning a blog post, that’s Google telling you the intent, and arguing with it rarely works. This single check, done consistently, saves more wasted content than any other step in this process.

Step 1: Build your seed list

Seed keywords are the small set of broad terms that describe what your business does, before any expansion. If you run a bakery, your seeds might be “custom cakes,” “wedding cakes,” “cupcakes,” “bakery near me.” Keep this list short, five to fifteen terms, covering your core products or services in the plainest language possible. Don’t try to be clever or comprehensive here. The seed list exists purely to feed the next step.

Step 2: Expand the list using a keyword tool

Take each seed term and run it through a keyword research tool. Google Keyword Planner is free and a reasonable starting point, though it’s built primarily for advertisers and sometimes groups volume data into broad ranges rather than exact numbers. Ahrefs, Semrush and Ubersuggest all offer more granular data and better keyword suggestion features, at a monthly cost.

Each tool will show you related terms, questions people ask around that seed, and rough search volume. Export or copy everything relevant, even terms that don’t look immediately obviously valuable. You’ll filter aggressively in a later step, so the goal right now is coverage, not precision.

Use the “People also ask” and autocomplete as free extras

Beyond dedicated tools, Google’s own “People also ask” boxes and autocomplete suggestions, what appears as you start typing a search, are free sources of real, current search behavior. Type your seed terms into Google’s search bar and note what autocomplete suggests. Search the term itself and note the “People also ask” questions that appear. These come directly from real search patterns and often surface long-tail phrasing that keyword tools miss entirely.

Step 3: Add question-based keywords deliberately

Beginners often collect statement-style keywords (“running shoes for flat feet”) and forget question-style ones (“what running shoes are best for flat feet,” “do I need special shoes for flat feet”). Questions matter for two reasons: they map cleanly onto FAQ sections and featured snippets, and they tend to have clearer, more specific intent than shorter phrase keywords, which makes them easier to write genuinely useful content against. Most keyword tools have a dedicated “questions” filter or report. Use it.

Step 4: Check search volume, but don’t worship it

Search volume tells you roughly how often a term is searched, usually shown as a monthly average. It’s a useful filter, but it’s not the only number that matters, and different tools frequently disagree with each other by a meaningful margin because they estimate from different data sources. Treat volume as directional: is this term searched by dozens of people a month, or thousands, not as a precise figure you can do exact math with.

For a brand-new site, I generally steer beginners away from chasing the highest-volume term in a topic first. A term with 200 monthly searches and low competition can realistically rank within months. A term with 20,000 monthly searches and a page full of established competitors might take years for a new site to compete for, if it happens at all.

Step 5: Check keyword difficulty and who’s actually ranking

Most keyword tools show a difficulty score, an estimate of how hard a term is to rank for based on the authority of sites currently ranking. Treat the score itself loosely; each tool calculates it differently. What’s more useful is actually looking at the top ten results for a term: are they large, established brands, or smaller, comparable sites to yours? If every result on page one is a major publication or a site with thousands of backlinks, that’s a realistic signal a new site will struggle there regardless of what the difficulty score says.

Step 6: Separate head terms from long-tail keywords

A head term is short and broad, “running shoes.” A long-tail keyword is longer and more specific, “best running shoes for flat feet under $100.” Individually, long-tail terms look unimpressive in a keyword tool, sometimes ten or twenty searches a month. Collectively, on most established sites, they bring in more total traffic than the handful of head terms, and they convert at a noticeably higher rate because the intent is so much more specific.

For a beginner’s first keyword list, I’d deliberately weight it toward long-tail terms. They’re realistically winnable sooner, and ranking for several of them builds the topical relevance that eventually helps with the harder head terms too.

Step 7: Group everything by topic, not by exact phrase

Once your raw list is long, and it should be, often hundreds of terms even for a small business, the next step is grouping, not filtering. Look for clusters of keywords that share the same underlying intent and could realistically be answered by one well-built page. “Best running shoes for beginners,” “running shoes for new runners,” and “what shoes should a beginner runner buy” are different phrasings of essentially the same question. They belong together, feeding one page, not three competing ones.

This grouping step is where a raw export from a keyword tool turns into an actual content plan. I’d rather see a beginner end up with fifteen well-defined topic groups than three hundred ungrouped keyword rows nobody will ever act on.

Step 8: Map each group to a page

For every topic group, decide whether it maps to a page that already exists on your site or a page you need to create. Write this mapping down somewhere you’ll actually reference later, a spreadsheet is fine. Skipping this step is how sites end up with two different pages unintentionally targeting the same topic, a problem called keyword cannibalization, where both pages can end up suppressing each other’s rankings instead of either one performing well.

Step 9: Prioritize using three factors together

With your grouped, mapped list in front of you, prioritize using relevance to your actual business first, realistic competition for your site’s current authority second, and volume third. A highly relevant, low-competition topic with modest volume is usually a better first target for a new site than a high-volume topic that’s only loosely related to what you actually offer, or one that’s realistically out of reach against established competition.

A mistake I see beginners make constantly

Building an enormous keyword list, hundreds or thousands of terms, and never getting past the spreadsheet stage. A keyword list only has value once it turns into actual pages. I’d rather see a beginner do a smaller, rougher version of this process and start publishing against ten well-chosen topics than spend a month perfecting a keyword spreadsheet that never becomes content. The list is a means, not the deliverable.

What keyword research can’t tell you

A keyword tool can tell you roughly how often a term is searched and what else people search alongside it. It can’t tell you whether you can actually write something better than what’s currently ranking, and it can’t guarantee a ranking position no matter how well the term matches your business. Treat keyword research as the input to a decision, not a promise about the outcome.

Where this fits in the bigger picture

Keyword research is step one of the process laid out in the full SEO for beginners guide, which covers how these keywords translate into on-page structure, technical fundamentals, and measuring whether any of it is actually working. If you want a structured, hands-on version of this entire process, including feedback on your actual keyword list, that’s what the SEO Masterclass walks through in detail.