How to Do Keyword Research (Step by Step)

To do keyword research, find the terms your audience actually searches, judge each one by intent and how winnable it is, then map every chosen keyword to a single page built to satisfy it. The goal isn’t a big list of high-volume words — it’s a prioritized plan of terms you can realistically rank for that reach the right people. Here’s the process, step by step.

This guide walks through discovery, intent, winnability, prioritizing, and mapping — the same process behind content that actually ranks.

Step 1 — Start with seed topics

Begin with the broad topics your business is about — the handful of subjects a customer would search around. For a web designer that might be “website cost,” “website redesign,” “web design for [industry].” These seeds are the starting points you’ll expand into real keywords. Don’t overthink it; list the themes your ideal customer cares about.

Step 2 — Expand into real search terms

Now turn seeds into the actual phrases people type. Sources that reveal real demand:

• Google autocomplete — start typing a seed and note the suggestions.
• People Also Ask — the related questions Google shows.
• Related searches — the block at the bottom of the results.
• Keyword tools — Ahrefs, Semrush, Google Keyword Planner for volume and variations.
• Competitor pages — what terms are competitors targeting?

You’re looking for the real questions and phrases, including specific long-tail ones, that your audience uses.

Step 3 — Classify the intent of each term

For every candidate keyword, ask what the searcher actually wants — the search intent:

• Informational — they want to learn (“what is schema markup”).
• Commercial — they’re comparing (“best web designers”).
• Transactional — they’re ready to act (“hire a web designer”).
• Navigational — they want a specific site.

The fastest way to confirm intent: search the term and look at what’s ranking. The format of the top results tells you the intent Google has already validated.

Step 4 — Assess winnability (not just volume)

This is where good research beats bad. Don’t chase the biggest numbers — assess whether you can realistically rank:

• Check keyword difficulty as a rough filter.
• Then look at the actual SERP — who ranks, how strong are they, how many have the exact term in their title? A low-difficulty score with entrenched, authoritative competition is not actually winnable.
• Match difficulty to your site’s strength — a newer site should target genuinely winnable long-tail terms, not impossible head terms.

Step 5 — Prioritize by intent, winnability, and value

Score your candidates on three things together: does the term reflect someone valuable to your business, can you realistically rank for it, and does ranking matter commercially? A specific, winnable, high-intent term usually beats a huge, unwinnable one. Volume is one input — not the goal.

Step 6 — Map each keyword to one page

Assign every priority keyword to exactly one page, so no two pages compete for the same term (which causes keyword cannibalization). This map becomes your content plan: pillars for competitive money terms, supporting articles for informational and long-tail terms, all interlinked with internal links. One keyword, one page, one job.

Step 7 — Revisit and expand

Keyword research isn’t one-and-done. As you publish and track performance in Search Console, you’ll discover new terms you’re nearly ranking for, gaps to fill, and questions to answer. Feed those back into the map. In the AI era, also watch the questions people ask AI tools — the same demand-and-intent skill applies to both search boxes and AI prompts.

Do this properly and every page you write has a real, winnable target and a clear purpose — which is what makes content rank. (SEO consulting builds content on a real keyword map, not guesswork.)

Frequently asked questions

How do you do keyword research step by step? Start with broad seed topics your business is about, expand them into real search terms using autocomplete, People Also Ask, related searches, and keyword tools, classify each term’s intent, assess how winnable it is by checking difficulty and the actual results, prioritize by intent, winnability, and business value, then map each chosen keyword to a single page. Revisit and expand as you track performance.

What is the first step in keyword research? The first step is identifying seed topics — the broad subjects your business is about that customers would search around. These starting themes get expanded into specific keywords in later steps. You don’t need tools yet; just list the topics your ideal customer cares about, then build outward from there using research sources to find the actual phrases people type.

Do I need paid tools for keyword research? Not necessarily. Paid tools like Ahrefs and Semrush provide helpful volume and difficulty data, but you can do meaningful research with free sources: Google autocomplete, People Also Ask boxes, related searches, Google Keyword Planner, and simply examining the actual search results for a term. The most reliable check — looking at what already ranks — is free and reveals both intent and competitiveness.

How do I know if I can rank for a keyword? Check keyword difficulty as a rough filter, then look at the actual search results — who currently ranks, how authoritative they are, and how many have the exact keyword in their title. A low difficulty score doesn’t guarantee winnability if strong competitors dominate. Match the keyword’s competitiveness to your site’s current strength; newer sites should target genuinely winnable long-tail terms first.

What matters more, search volume or intent? Intent and winnability generally matter more than raw volume. The highest-volume keyword is often the wrong first target — it may be impossibly competitive or attract searchers who won’t convert. It’s better to target specific, winnable, high-intent terms you can realistically rank for and that reach valuable customers. Treat volume as one useful input, not the goal of keyword research.

How many keywords should I target per page? Target one primary keyword per page, with the page naturally covering closely related terms and questions around it. Assigning one main keyword per page prevents keyword cannibalization, where multiple pages compete for the same term and dilute each other. A well-optimized page can rank for many related variations, but it should be built around a single clear primary target.

How often should I do keyword research? Keyword research is ongoing, not a one-time task. Revisit it regularly as you publish content and track performance in Search Console, where you’ll spot new terms you’re nearly ranking for, gaps to fill, and emerging questions. Markets and search behavior shift over time, so periodically refreshing your research keeps your content plan aligned with real, current demand.

Written by Bryan Collins, SEO & AEO strategist. Want a content plan built on real demand? See how I work with clients on SEO or run a free site audit.