Keyword research is the process of finding the words and phrases people type into search engines, then deciding which ones your site should target. It is the planning stage of SEO: get it wrong and every article, link and technical fix is pointed at the wrong goal.
Good research answers three questions. What are people searching for? Which of those searches could turn into customers? And which can you realistically rank for?
Step 1: Brainstorm seed keywords
Write down the products, services and problems your business deals with, in your customers’ words rather than your internal jargon. A dentist might say “dental implants”, but patients also search “replace missing tooth” and “implant cost UK”. Check your sales emails, reviews and support tickets for phrasing.
Step 2: Expand the list with tools
Paste the seeds into a keyword tool. Free options include Google Keyword Planner, Google’s autocomplete and the People Also Ask box. Paid tools such as Ahrefs, Semrush and Moz show volume, difficulty and the pages already ranking.
Your own data is the richest source of all. The Performance report in Google Search Console shows queries you already appear for, often in positions 8 to 20 where a small push can bring big gains.
Step 3: Judge volume and value
Monthly search volume tells you how many people search a term, but not how valuable they are. “Accountant” has huge volume and mixed intent; “accountant for limited company Bristol” has tiny volume and a buyer behind almost every search.
That is why long-tail keywords often outperform head terms for small and new sites: less competition, clearer intent and higher conversion rates.
Step 4: Check difficulty
Tools give a keyword difficulty score based mainly on the links pointing at the top results. Use it as a filter, then look at the results page yourself. If forums, thin pages or small sites rank, you have an opening, whatever the score says.
Step 5: Identify search intent
Search the term and study page one. Informational searches show guides, commercial searches show comparisons, transactional searches show product pages and local searches show the map pack. Matching search intent is not optional: you cannot rank a sales page for a query Google treats as informational.
Step 6: Group keywords into topics
Many keywords share the same intent and the same results. “How to fix a leaking tap”, “leaking tap repair” and “dripping tap fix” belong on one page. Grouping prevents keyword cannibalisation and helps each page rank for dozens of variations.
Groups then roll up into clusters around a pillar page, which is how you build topical authority over time.
Step 7: Map keywords to pages
Create a simple spreadsheet: keyword group, primary keyword, intent, volume, difficulty, target URL and status. Every important page gets one primary group, and every group gets one page. This map becomes your content plan.
Keyword research for AI search
People ask AI assistants longer, conversational questions. Mine forums, Reddit and People Also Ask for natural phrasing, and make sure your pages answer those questions clearly. Our guide to generative engine optimisation covers this in more depth.
Common mistakes
- Chasing volume over relevance.
- Targeting one keyword per page instead of a topic.
- Ignoring the actual results page.
- Never revisiting the research as the market changes.
Frequently asked questions
How many keywords should a page target?
One primary keyword plus the close variants and questions that share its intent, often dozens.
Are zero-volume keywords worth targeting?
Often yes. Tools underreport niche B2B and local terms, and a single sale can justify the page.
Related terms
See semantic SEO, SEO content writing and the SEO encyclopedia.
