How to Do Keyword Research (Free Method That Actually Works)
Here is how to do keyword research without paying for a single tool: list the questions your customers actually ask, expand that list using Google's own free data (autocomplete, People Also Ask, Search Console, Keyword Planner and Trends), then sort what you find by how likely you are to rank rather than by search volume. That's the whole method. Everything below is detail on doing each step well.
Most people get this backwards. They open a tool, type one word, sort by highest volume, and pick terms they have no realistic chance of ranking for. Keyword research is not a search for the biggest number. It's a search for the gap between what people want and what the current results actually give them.
What keyword research is really for
A keyword is shorthand for an intention. Someone typing "how to do keyword research" wants a method. Someone typing "keyword research tool pricing" wants to compare and buy. Same topic, completely different pages required.
So before the tools, answer three questions:
- Who is searching? A solo shop owner and an in-house SEO lead use different words for the same problem.
- What do they want at that moment? To learn, to compare, to buy, or to find a specific site.
- What would a genuinely satisfying answer look like? If you can't beat what's already on page one, pick a different term.
If you're new to the wider subject, our plain-English guide to what SEO is covers the fundamentals this sits inside.
Step 1: Build a seed list before you touch a tool
Open a blank document and write down 10–20 terms from your own head. Pull them from real sources rather than imagination:
- Questions customers email or message you
- The words people use on sales calls — their words, not your industry's jargon
- Your service or product pages, and your competitors'
- Subreddits, forums, groups and Q&A threads in your niche
- Review sites — the phrases in complaints are especially useful
This is the step almost everyone skips, and it's the one that makes the rest work. Tools expand a list. They can't invent the right starting point.
Step 2: Expand with free Google data
Google hands you a large amount of keyword intelligence at no cost. You just have to collect it manually.
Autocomplete
Type your seed term into Google and note every suggestion. Then type it with a letter after it — "keyword research a", "keyword research b" — and note those too. These suggestions are drawn from real query patterns, so they reflect how people actually phrase things.
People Also Ask
Search your seed term and expand the "People Also Ask" box. Each answer you open loads more questions. Twenty minutes here typically produces dozens of genuine questions, each one a potential H2 in your article or a page of its own.
Related searches
Scroll to the bottom of the results page. The related-searches block shows what people search next — useful for mapping a topic rather than a single term.
Google Trends
Trends doesn't give absolute volume, but it gives something more useful for planning: direction and seasonality. Compare two terms side by side to see which is growing, and check the 12-month curve so you publish seasonal content before the peak, not during it.
Google Search Console
If your site has any traffic at all, this is your single best free source, because it's your data rather than an estimate. Go to Performance → Queries and look for:
- Terms ranking in positions 5–20 with decent impressions — the fastest wins, since improving an existing page is far easier than ranking a new one
- Queries with high impressions but a low click rate — usually a weak title or a mismatch between the page and the intent
- Queries you didn't intend to rank for — often a topic worth its own page
Google Keyword Planner
Keyword Planner is free, with one caveat worth knowing: accounts without active ad spend generally see volume in broad ranges (10, 100, 1K–10K) rather than exact figures. For most planning that's fine — you're separating "plenty of demand" from "almost none", not forecasting traffic to two decimal places. Treat every volume number, from any tool, as directional.
Free tools worth adding
These fill the gaps Google's own tools leave. Free tiers change often, so confirm current limits before you rely on one.
| Tool | Best for | Typical limitation |
|---|---|---|
| Google Search Console | Your real queries and positions | Only terms you already rank for |
| Google Keyword Planner | Demand size and seasonality | Broad ranges without ad spend |
| Google Trends | Direction and comparisons | Relative, not absolute, numbers |
| AnswerThePublic | Question-shaped keywords | Few free searches per day |
| Ubersuggest | Quick difficulty sanity-check | Few free queries per day |
| Bing Webmaster Tools | A second free keyword dataset | Smaller search share than Google |
Step 3: Judge difficulty without a paid tool
Paid tools give you a difficulty score. You can approximate one by looking at page one and asking honest questions:
- Who's ranking? Ten major brands and publishers means a hard term for a small site. A mix of small sites and forum threads means an opening.
- How good is the content? If the top results are thin, outdated, or don't actually answer the question, that's your gap.
- Does the format match? If page one is all comparison tables and you planned an essay, Google has already told you what it wants.
- How old are the results? Several four-year-old posts on a fast-moving topic is an invitation.
As a rule, a newer site should aim at specific, longer phrases — four to seven words, clearly defined intent — and work up. "Keyword research" is a fight. "How to do keyword research without paid tools" is a page you can win.
Step 4: Group keywords into pages, not one keyword per page
This is where current keyword research differs most from the old playbook. Search engines now understand topics, not just strings, so ten near-identical pages targeting ten variations of one phrase compete with each other and satisfy nobody.
Cluster instead. Put every term that would be answered by the same page into one group, pick the clearest phrase in the group as the primary keyword, and let the rest become section headings and natural mentions. One thorough page usually outranks five thin ones.
Step 5: Account for AI answers
AI-generated summaries now appear on a meaningful share of results pages, and they change what a good keyword looks like. Two practical consequences:
Queries are getting longer and more conversational. People increasingly ask full questions rather than typing two words, then follow up on the answer. That makes question-shaped, specific phrases more valuable than they used to be — and it makes People Also Ask an even better research source.
Simple factual queries convert to clicks less often. If the entire answer fits in two sentences, a summary may deliver it and the searcher never visits a page. Weight your list toward topics where the reader needs depth: a process, a judgement call, a comparison, or your specific experience — things a summary can't fully replace.
The structural response is the same either way: answer the core question clearly in your first two paragraphs, then go deeper than a summary can. That's what earns both the featured snippet and the click. Our guide to getting your first 1,000 website visitors goes further into turning rankings into actual visits.
Step 6: Build a simple tracker
A spreadsheet is enough. Columns: keyword, intent, rough volume, difficulty guess, target page, status, current position. Sort by a simple priority — decent demand, beatable competition, clearly relevant to what you sell.
Revisit it quarterly. Keyword research is not a one-time project; positions move, competitors publish, and Search Console keeps handing you terms you never thought to target.
Common mistakes to avoid
- Chasing volume alone. A term with 200 searches and buying intent beats one with 20,000 and none.
- Ignoring what already ranks. Page one is a free brief on what Google thinks the searcher wants.
- Targeting terms you can't win. Ambition is fine, but a first-year site aiming only at head terms is a plan to get no traffic.
- Keyword stuffing. Use the phrase where it reads naturally — title, first paragraph, a heading — then write for the person.
- Never checking results. Six months of Search Console data will teach you more about your niche than any tool.
Frequently asked questions
How long does keyword research take?
For a single article, 30–60 minutes is usually enough. For a full site plan, expect a few focused sessions spread over a week.
Do I need a paid tool?
Not to start. Paid tools save time and add competitor data, which matters at scale. If you're publishing a few posts a month, free sources cover it.
How many keywords should one page target?
One primary phrase, plus every close variation and question that belongs to the same intent — often ten to thirty related terms handled naturally through headings and body copy.
Is keyword research still worth it with AI search?
Yes, though the emphasis has shifted. You're researching questions and intent more than exact strings, and choosing topics where a short summary isn't a complete answer.
What's the difference between short-tail and long-tail keywords?
Short-tail terms are one or two words with broad demand and heavy competition. Long-tail terms are longer, more specific phrases with less demand each but clearer intent and a far better chance of ranking.
Start with what you already have
If you take one thing from this: open Search Console, sort your queries by impressions, and find the ones sitting in positions 5–20. Those pages already have Google's attention. Improving them is faster and more reliable than starting from zero.
If you'd rather have someone map the keyword strategy for your site — which terms are realistically winnable, and which pages should target them — get in touch and we can talk it through.