A paid keyword tool gives you two things you cannot get free: search volume estimates for markets you have no presence in, and a backlink index. Everything else in keyword research is available without paying anything.
And for a small site, the free sources are frequently better — because they are your own measured data and your own customers' actual questions, rather than a model of a market.
What follows is the complete method, the honest account of what you are giving up, and a routine that takes about half an hour a week.
The seven sources
Ordered by value, not by convenience.
1. Search Console — your own queries
The best keyword tool available to you, and it is free because it is yours.
Performance → Queries shows what you already appear for. Three views worth having:
- Positions 8 to 20, sorted by impressions. Google considers you relevant; you are below where clicks happen. This is the opportunity list
- High impressions, low clicks. You are being shown and not chosen — usually a title, a snippet, or a near-miss on intent
- Filtered to question words — how, why, what, can, does. The long tail you are already reaching
Its documented limits, which you should know rather than discover:
- "Tables in the performance reports omit rare queries to protect user privacy" — the rarest tail is invisible
- "Our tables can show a maximum of 1,000 rows, so some rows might be omitted" — filter rather than scroll
- Position is an average of a topmost result, personalised per searcher. A band, not a rank
Reading it properly is covered in Google Search Console reports, read properly.
2. The questions you are actually asked
Support email, sales calls, comments, DMs, the "quick question" at the end of a call.
These are real queries in their original wording, and they are proof of demand no tool can offer: someone needed the answer enough to ask a person.
Keep a running note. Add to it as questions arrive rather than trying to remember them later.
3. The results page itself
Free, and it is Google showing you real related demand.
- "People also ask" — expands as you open entries, and goes deep
- Related searches at the bottom of the page
- Autocomplete, including the variations that appear as you add words
Read them for phrasing as much as topic. How people ask is not how marketers write, and matching their words is most of the work.
4. Google Trends
Free, and it does something no paid tool does better: relative interest over time.
Use it for:
- Direction — is this subject growing or fading?
- Seasonality — when does interest peak?
- Comparison between two phrasings, which tells you which one people actually use
Not for absolute volume, which it does not give. It is a shape, not a number, and the shape is often more decision-relevant than the number.
5. Keyword Planner
Free with a Google Ads account, and usable without spending — with caveats you should hold firmly.
Google documents that the figures are rounded: "Your search volume statistics are rounded. This means that when you get keyword ideas for multiple locations, the search volumes might not add up as you'd expect." They are averages across a chosen range, location and network, and Google notes that seasonality and current events move them constantly.
Use it for discovering related terms and rough relative scale. Do not use it to forecast traffic, and do not treat a reported difference between two small numbers as real.
6. Where people ask each other
Forums, communities, Q&A sites, subreddits, the comments under competitors' articles.
Look for the question that got no good answer, or an answer that has since gone stale. Real demand plus poor supply is the clearest opportunity signal there is, and it requires no subscription.
It is also where you find the wording people use when they are frustrated, which is usually the exact phrasing they search.
7. Your competitors' pages, read properly
Not their keyword data — their actual pages.
Read five of them properly, not with a tool. What do they all cover? What do they all miss? What do they get wrong? What questions do their comments ask that the article did not answer?
This is slower than an export and it produces better decisions, because you finish it knowing what you would add — which is the gate every page has to pass.
The routine
Seven steps. About half an hour a week once it is running.
Step 1 — Export Search Console monthly
Queries and Pages, last three months, saved somewhere you keep them. Month-on-month comparison is where the useful movement shows up.
Step 2 — Pull the 8-to-20 band
Sort, filter, and read the queries. Anything you did not expect to appear for is worth a note.
Step 3 — Add this week's real questions
From email, support and calls. In the asker's words.
Step 4 — Take the three most promising and read their results pages
Format of the top ten, result features, and the top three read properly. Method in search intent.
This is the step that decides whether to proceed, and it is why the list stays at three rather than thirty.
Step 5 — Check your own site first
site:yourdomain.com "the phrase" before writing anything. Thirty seconds, and it prevents the problem described in keyword cannibalisation.
Step 6 — Group and decide
One page per question, not per phrasing. Then decide: improve an existing page, or create a new one.
Improve wins more often than people expect, and it produces results sooner.
Step 7 — Record the decision and check it later
Query, decision, date. Then look at Search Console three months on and see whether the page did what you thought it would.
This is the step everyone skips, and it is the only one that makes the method improve over time.
What you actually give up
Stated honestly, because the free-method genre tends not to.
Volume for markets you have no presence in. If you are entering a subject where you have no impressions, Search Console has nothing to tell you and Keyword Planner's ranges are all you have.
Competitor keyword data. You cannot see what queries a competitor ranks for, only what pages they have published. You lose speed here, not accuracy — reading their pages tells you the same things more slowly.
A backlink index. Genuinely unavailable free, and the real product of the paid tools — the argument in Ahrefs vs Semrush vs free alternatives.
Bulk analysis. Comparing fifty domains or auditing ten thousand URLs is not a manual job.
Difficulty scores. These are third-party estimates rather than measurements, so what you lose is a convenient proxy rather than a fact.
What you do not give up: knowing what your own site ranks for, what your customers ask, what the results pages demand, whether a subject is growing, and what competitors have failed to cover. That is most of keyword research.
Who should pay
Not a hedge — a list.
Pay if you are doing this for clients. The tooling is part of the deliverable and the reporting is the product.
Pay if you are entering a market where you have no presence at all, and need volume estimates before committing to a content programme.
Pay if you are doing active link acquisition, where the backlink index is the whole point.
Pay if manual work is the bottleneck — many sites, many competitors, many URLs.
Do not pay if you have a site with Search Console data, customers who ask questions, and time to read results pages. That describes most small businesses, and the money is better spent on producing the pages.
Frequently asked questions
Can you do keyword research without paid tools?
Yes, for most small sites completely. Search Console shows what you already rank for, customers supply real questions, results pages show related demand, and Trends shows direction. Paid tools add volume estimates for unfamiliar markets and a backlink index.
What is the best free keyword research tool?
Search Console, because it reports your actual performance rather than an estimate of a market, and no competitor can see it. Google Trends and Keyword Planner cover direction and rough scale.
Is Google Keyword Planner accurate?
It gives rounded averages, and Google's own documentation says the figures are rounded and may not add up as expected across locations. Use it for discovering related terms and rough relative scale, not for forecasting traffic.
How do I find keywords without a subscription?
Export Search Console queries and read the positions 8-to-20 band, collect the questions customers actually ask, expand "people also ask" and related searches, check direction in Trends, and read five competitor pages properly.
Is Google Trends useful for keyword research?
Yes, for relative interest over time — direction, seasonality, and which of two phrasings people actually use. It does not give absolute volume, so treat it as a shape rather than a number.
What do paid SEO tools give you that free ones don't?
Search volume estimates for markets where you have no presence, competitor keyword data, bulk analysis across many domains or URLs, and a backlink index. The backlink index is the one with no free equivalent.
How often should I do keyword research?
A monthly Search Console export with a weekly half-hour routine is enough for most sites. The step that matters most is recording what you decided and checking three months later whether it worked.
Why doesn't Search Console show all of my keywords?
Google omits rare queries from performance tables to protect user privacy, and tables cap at 1,000 rows. Both are documented, and it is why what you see is a sample of the tail rather than all of it.
What to do next
Open Search Console, export the last three months of queries, and filter to rows where average position is between 8 and 20.
That export is a keyword research report about your own site, generated from real data, that no competitor can see and that cost nothing. Read it before you consider paying for anything.
Related guides
- Keyword research that starts with your own data — the full method
- Search intent: the four types and what they change — reading the results page
- Long-tail keywords — where the free sources are strongest
- Keyword cannibalisation — the check before writing
- Google Search Console reports, read properly — the primary source
- Ahrefs vs Semrush vs free alternatives — if you decide to pay
- Refreshing old blog posts — where improve-work happens
The short version
- Export Search Console monthlyQueries and Pages, last three months, kept somewhere you can compare month on month. The movement between exports is where the useful signal is.
- Pull the 8-to-20 position bandSort by impressions and read the queries themselves. Anything you did not expect to appear for is worth a note.
- Add this week's real questionsFrom email, support and calls, written in the asker's words rather than rewritten into keyword phrasing.
- Take the three most promising and read their results pagesFormat of the top ten, the result features, and the top three read properly. This step decides whether to proceed.
- Check your own site firstA site: search on the phrase before writing anything. Thirty seconds, and it prevents two of your own pages competing.
- Group and decideOne page per question, not per phrasing. Then choose: improve an existing page or create a new one. Improve wins more often than expected.
- Record the decision and check it laterQuery, decision, date u2014 then look at Search Console three months on. This is the step everyone skips and the only one that makes the method improve.
Free: The 60-Minute Email Authentication Fix
A no-fluff checklist to set up SPF, DKIM & DMARC correctly and pass Gmail & Yahoo's sender requirements.
Muhammad Basim has worked in digital marketing since 2013, focused on email deliverability and AI-assisted content production. He is the author of The Email Deliverability Playbook and The Email Copywriting Playbook.
Related Articles
Email Deliverability: Why Authenticated Emails Still Land in Spam
Email deliverability is whether your message reaches the inbox rather than the spam folder, and it is decided by three factors: technical infrastructure, list quality, and sending behaviour. Authentication is one item inside the first factor. Getting it right is necessary, and it is nowhere near sufficient. That gap explains the most common complaint in […]
Linkable Assets: What Actually Earns Links
A link is a citation, and a citation requires that someone writing about your subject needed you to make their point. That is the whole mechanism, and it is the reason most "linkable content" earns nothing. The test, before you build anything: Could a writer covering this topic finish their sentence without referencing you? If […]
Link Outreach Emails That Get Replies
Outreach fails for exactly two reasons, and only one of them gets written about. The first is that the email was not worth replying to. That is the reason every guide addresses, and the advice — personalise, be brief, offer value — is correct and insufficient. The second is that the email never arrived. Nobody […]