Keyword Research for SEO

What is keyword research for SEO?

Keyword research for SEO is the process of finding the words and questions real people type into search engines, then choosing which ones your site should target based on search volume, difficulty and intent. It is the foundation of SEO: it tells you what to write, which pages to build, and which terms are actually worth competing for.

Done well, keyword research turns guesswork into a plan. Instead of writing whatever you feel like, you discover the exact phrases your customers use, see how many people search them each month, judge how hard they are to rank for, and map each one to a page. The output is a prioritized list of topics that can realistically rank and drive qualified traffic.

In 2026 the job has expanded. You are no longer just optimizing for ten blue links. You are also researching the questions and entities that AI search engines like Google AI Overviews, ChatGPT and Perplexity pull from when they write an answer. The good news: the same research that wins Google rankings is the raw material for getting cited by AI, and we cover both on this page.

Why keyword research matters

Keyword research matters because it aligns your content with real demand. Without it you risk writing pages no one searches for, chasing terms too competitive to win, or missing the easy long-tail queries that convert. With it, every page you publish has a known audience, a measurable opportunity and a clear job to do.

It also drives the rest of your strategy. The keywords you find shape your site structure, your internal linking, and which topics deserve a deep pillar page versus a quick supporting article. Skip the research and you are building on sand.

The numbers are blunt. Ahrefs found that 96.55% of pages get no Google traffic at all, most often because they target the wrong terms or none. Meanwhile almost 93% of keywords get fewer than ten searches a month, so the winnable opportunities are usually the specific long-tail phrases most people skip. And because 15% of daily searches are brand new, the work never really ends. Keyword research is how you point your effort at demand that actually exists.

What you get from good keyword research

• Demand validation: proof that people actually search for a topic before you invest in it.
• Realistic targets: keyword difficulty tells you what you can win now versus later.
• Intent match: you write the content type searchers (and AI engines) expect for each query.
• A content map: every keyword maps to a page, so you avoid thin or duplicate content.

Keep it simple: the two types of keywords

Keyword research used to be a precise science of chasing exact-match search volume. AI search engines have made it fuzzier, because they read topics and questions rather than single phrases, so obsessing over one keyword and its exact volume matters less than it used to. That is why we simplify it. For most businesses, keyword research comes down to finding two types of keyword.

1. Informational keywords

Informational keywords are the questions and topics people search to learn something, like "how often should you service a hot water system" or "what is generative engine optimization". They carry research intent, so they map to your blog posts, guides and FAQs. This is how you earn trust, answer the sub-questions AI engines fan out into, and reach people at the top of the funnel before they are ready to buy.

2. Transactional keywords

Transactional keywords carry buying or action intent, like "emergency plumber near me", "repipe service cost" or "book a dental checkup". They map to your money pages: service pages, product pages and location pages. Fewer people search them, but the ones who do are close to a decision, so these are the terms that turn into calls and sales.

Sort your keywords into those two buckets first and the rest of the work gets obvious: informational terms tell you what content to write, transactional terms tell you what belongs on your service and product pages. This is really just search intent applied, and it is enough to build a plan that covers both the people learning and the people buying.

What about LSI keywords?

LSI keywords (latent semantic indexing keywords) are the related, contextual terms that naturally surround your main topic. For a page about repiping, that includes "copper pipes", "water pressure", "corrosion", "insurance" and "plumbing permit". They are not synonyms to stuff in, they are the concepts a genuinely complete answer would mention. The name is a throwback to older search technology, and Google does not literally use "LSI", but the idea still holds: cover the related terms and entities around your keyword so search engines and AI engines can see that your page fully covers the topic. The easiest way to find them is to look at what already ranks and at the People Also Ask box, then make sure your page addresses those points naturally instead of repeating your exact keyword over and over.

My favorite place to start: your competitors

One of my favorite ways to start keyword research is to skip the blank page and look at competitors first. Tools like DataWise or Semrush let you drop in a competitor's domain and see the exact keywords they already rank for.

The beauty of this is that those keywords are already vetted. A competitor ranking for a term is proof that it has real search demand, that it is winnable in your niche, and that it attracts the kind of customer you both want. Someone has effectively done the validation for you, so why start from scratch? Pull their ranking keywords, sort them into your informational and transactional buckets, and you have a working shortlist in minutes.

A simple way to run it: start with two or three close competitors, look for the terms they rank for but you do not (those are your content gaps), and prioritize the ones that match a page you can realistically create or improve. You are not copying them, you are using their proven keywords as a starting list, then adding your own experience and angle so your page becomes the better answer.

How to do keyword research for SEO (step by step)

Here is the core process in five steps, and the full walkthrough lives in how to do keyword research.

Step 1: Brainstorm seed keywords

Start with seed topics you already know: the products, services and problems your business solves. Expand them using free sources like Google autocomplete, People Also Ask, related searches, Reddit threads and YouTube suggestions. These reveal the language your audience actually uses.

Step 2: Expand with a keyword tool

Feed your seeds into a keyword research tool. Free options like Google Keyword Planner, Google Search Console and Google Trends get you started; tools like Ahrefs, Semrush, Ubersuggest and our own DataWise expand each seed into hundreds of related terms and questions, with metrics attached.

Step 3: Read the metrics

For each keyword, look at search volume (monthly searches), keyword difficulty (how hard it is to rank), cost per click (commercial value) and search intent. Understanding these metrics is what separates a random list from a strategy. Classify the intent (the why behind each query) by reading the SERP: the kind of pages Google already ranks tells you what format wins.

Step 4: Prioritize keywords

You cannot target everything at once. Prioritize by balancing search volume, keyword difficulty, intent and business value. New sites should lean on long-tail keywords: lower competition, clearer intent and higher conversion, even if each has smaller volume.

Step 5: Cluster and map keywords to pages

Group related keywords that share intent into clusters, then map each cluster to one page. This builds topical authority and prevents two of your own pages from competing. You can do this clustering with AI in minutes: see our AI keyword research guide.

Keyword research for AI search and answer engines (GEO)

Keyword research still matters with AI search, but the unit of research shifts from single keywords to questions and entities. AI engines like Google AI Overviews, ChatGPT and Perplexity break a query into many sub-questions (called query fan-out) and assemble an answer from sources that cover each one clearly. Your job is to research and answer those sub-questions.

Practically, that means three additions to classic research: collect the full set of questions around a topic (not just the head term), map the entities and concepts an answer should mention, and write pages that answer each sub-question directly in the first sentence or two. This is the heart of AI SEO and getting cited by answer engines.

Using ChatGPT and AI tools for keyword research

You can use ChatGPT and other AI tools to speed up keyword research: brainstorming seed topics, expanding a seed into question lists, clustering terms by intent, and drafting content briefs. AI is excellent at the creative and organizing work that used to take hours in a spreadsheet.

The catch is data. Large language models do not know real search volume or current difficulty, so they will confidently invent numbers. The reliable workflow is AI for ideas and structure, then a real keyword tool to validate volume, difficulty and intent before you commit. We walk through prompts and the full process in AI keyword research.

Inside the community we use DataWise to do both halves in one place: AI-generated ideas and clusters, backed by live keyword data and intent tagging, with AI-Overview opportunity scoring so you can see which terms are worth chasing for citations. DataWise is free for members.