- Keyword research finds the words and questions people use when they need what you sell.
- Judge keywords on intent, value, demand and difficulty, not on volume alone.
- Group related keywords into topics and give each topic one page.
- Clear pages that answer a topic well rank in search and are easier for AI answers to quote.
Keyword research is how you find the words and questions people type, or say, when they need what you sell. It tells you what to write, which page should answer it, and which searches are worth the effort. Done well, it is less about lists of terms and more about understanding demand.
Keyword research, in plain terms
Every search starts with a need. Someone wants to learn something, compare options, find a place or buy. The words they use are clues to that need. Keyword research is the work of collecting those clues, sorting them by intent and value, and deciding which page on your site should answer each one.
Google says it plainly in its own starter guide: think about the words a user might search for to find your content, and use them naturally on the page.1 Keyword research is the structured version of that advice.
Why it still matters when AI answers the question
Search now happens in more places: classic results, AI Overviews, chat assistants, maps and marketplaces. The queries are longer and more conversational, but the job is the same. You need to know what people ask before you can be the answer.
What has changed is the unit of work. Single keywords matter less than topics and the questions around them. A page that answers a cluster of related questions clearly tends to rank in classic search and is easier for AI systems to quote.
The four kinds of search intent
Before you judge a keyword by its volume, judge it by what the searcher wants. Most queries fall into one of four groups.
Intent decides the page type. Informational queries want a guide. Commercial queries want a comparison or a service page with proof. Transactional queries want a clear offer and an easy next step. Matching the page to the intent is often worth more than finding a bigger keyword.
How to do keyword research, step by step
Start with what you sell and who buys it
List your products or services, the problems they solve and the people who buy them. Write down the words your customers use, not your internal names. Sales calls, support tickets, reviews and onboarding questions are the best sources here.
Collect seed terms and real questions
Turn that list into seed terms: short phrases that describe each offer. Then gather the questions around them from search suggestions, "People also ask" boxes, forums, and the queries already in your Search Console Performance report.3
Expand the list with a tool
Keyword tools show related terms and rough demand. Google Keyword Planner gives search volume ranges and related ideas from Google Ads data.2 Third-party tools add difficulty scores and the pages that already rank. Treat every number as a direction, not a fact: volumes are estimates and change by season and market.
Judge each keyword on four things
Volume alone is a poor guide. Score each keyword on the four questions below, and you will spot the ones worth doing first.
Group keywords into topics
Put keywords that share an intent into one group. One group usually becomes one page. This stops you writing five thin pages that compete with each other, and gives you one strong page that answers the whole topic.
Map every group to one page
Build a simple keyword map: the topic, its main keyword, the supporting questions, the intent, and the URL that owns it. If a page already exists, improve it. If not, add it to the content plan. Each URL should own one topic, and each topic should have one URL.
Turn the research into pages
Research only pays off when it changes a page. Use the main keyword in the title and the main heading, answer the core question near the top, and cover the supporting questions in clear sections. Write for the reader first: Google is explicit that content should be made for people, not for search engines.4
Common keyword research mistakes
Most mistakes come from treating keywords as a list instead of a map. Teams chase the biggest numbers and ignore intent, so traffic grows while leads do not. Several pages target the same term and split the results between them. Research is done once and never revisited, while the way people search keeps moving.
The fix is simple, if not always easy: fewer, better pages, each owning one topic, each matched to what the searcher wants, and a habit of checking what people actually searched for in Search Console.
The bottom line
Keyword research is how you learn the language of demand. Find the questions your buyers ask, judge them by intent and value, group them into topics, and give each topic one strong page. That approach works in classic search, and it is what makes your pages easy for AI answers to use.
Questions people ask
What is keyword research in simple words?
It is finding the words and questions people search for when they need what you offer, then deciding which page on your site should answer each one.
Is keyword research still useful with AI search?
Yes. People still ask questions, just in longer and more natural ways. Knowing those questions is how you build pages that rank and that AI answers can use.
How many keywords should one page target?
One main keyword plus the closely related questions around it. If two keywords need different answers, they need different pages.
Which tools do I need for keyword research?
Start with free ones: Google Search Console for the queries you already appear for, and Google Keyword Planner for related ideas and volume ranges. Paid SEO tools add difficulty scores and competitor data.
How often should keyword research be updated?
Review your keyword map every quarter, and whenever you launch a new product, enter a new market or see new queries in Search Console.