How to Do Keyword Research for Free (No Paid Tools Required)

Every article that ranks starts with the same question: what are people actually typing into Google? Keyword research is the work of answering that question before you write — and the common belief that you need a paid tool to do it is wrong. The free methods in this guide find the same keywords the paid tools do. They take longer, and they show less data per keyword, but for a small site choosing its next ten topics, they are enough.

The method below is the one I would use on a brand-new site with no budget. It has eight steps, it uses nothing but free tools, and it ends with a keyword map you can write from for months.

Step 1: start with seed keywords, not tools

Open a blank document and write down the topics your site covers, in your readers' words — not yours. A hosting site's seeds might be 'start a blog', 'cheap web hosting', 'WordPress backup', 'website speed'. A freelancer's seeds might be 'invoice template', 'client contract', 'find freelance clients'. Aim for ten to twenty seeds.

Then, for each seed, write down the questions a beginner would ask about it. 'How much does hosting cost?' 'How do I back up WordPress without a plugin?' 'What should a freelance contract include?' These questions are the raw material. Tools refine them; they do not replace them. If you skip this step and go straight to a tool, you will end up writing about what the tool suggests instead of what your readers need.

Step 2: mine Google autocomplete and People Also Ask

Type a seed into Google and stop before pressing Enter. The autocomplete suggestions are phrases real people have searched — Google shows them because they are popular. Add an underscore trick: type the seed followed by a letter ('how to back up WordPress a', '... b', '... c') and Google will show suggestions starting with each letter. It is slow, manual work, and it is one of the best free keyword sources that exists.

Then search the seed properly and read the 'People Also Ask' box. Each question expands into more questions when you click it, and the box keeps growing — click through ten or twelve and you have a list of real reader questions, phrased the way readers phrase them. Scroll to the bottom for 'Related searches' and collect those too. Every phrase you collect goes into your list with a note of where it came from.

Step 3: use Google Search Console — your own data

If your site has been live for even a month, Search Console is your best keyword tool, and it is free. Open the Performance report, sort by impressions, and look at the queries your pages already appear for. You will find three kinds of opportunity: queries where you rank on page one and could rank higher with a better article, queries where you rank on page two or three and a focused article could push you onto page one, and queries you never targeted at all that your pages are accidentally appearing for.

The second group is the cheapest win in SEO. A page ranking 12th for a query is one good rewrite away from the first page. Note these queries, then write or improve the matching content. This is also how you find the exact words your audience uses — Search Console shows their phrasing, not your guess at it. If you have not set Search Console up yet, our guide to Core Web Vitals explains why it matters; the setup itself takes minutes.

Step 4: check timing with Google Trends

A keyword with steady volume is different from one that spikes every November. Paste your candidate keywords into Google Trends and look at the five-year graph. Rising interest means a growing topic worth investing in. A sharp annual spike means a seasonal topic — plan the article a month before the spike, not during it. A long decline means the topic is dying, and your effort is better spent elsewhere. Trends also shows related queries and regional interest, which is useful if your audience is concentrated in one country.

Step 5: get volume estimates from free tools

Autocomplete and Search Console tell you what people search; they do not tell you how many. For that, use the free tiers:

Google Keyword Planner. Free with a Google Ads account (you never need to run an ad). It gives volume ranges rather than exact numbers, and it is built for advertisers, so treat the figures as rough. Still, it is Google's own data, and it is free.

AnswerThePublic. Enter a seed and it returns the questions, prepositions, and comparisons people search around it, visualised as a wheel. The free tier allows a few searches a day — enough for a small site planning a week of content.

Ubersuggest free tier. A few free searches a day with volume, difficulty estimates, and content ideas. The numbers are estimates, not measurements, but they are consistent enough to compare keywords against each other.

Keyword Surfer. A free Chrome extension that shows estimated volume directly in Google search results as you browse. It is the fastest way to sanity-check a keyword while doing step 2.

Use two of these, not all four. The point is a rough volume band — under 100, hundreds, thousands — not a precise number. Nobody's volume data is exact, paid or free.

Step 6: judge difficulty without paid metrics

Paid tools sell a difficulty score. You can approximate it for free in about two minutes per keyword. Search the keyword in Google and look at the top ten results. Ask four questions: Are the ranking pages from big authority sites, or from small sites like yours? Do the titles match the keyword exactly, or are they ranking by accident? Is the content thin or outdated — a 2019 article, a forum thread, a page that barely answers the question? And is there a featured snippet you could take with a better answer?

If the top results are small sites with thin, dated content, you can compete. If they are Wikipedia, the vendor's own documentation, and three major publications, move on — that keyword belongs to someone with more authority than you have today. This manual check is slower than a difficulty score and more honest than one, because you are looking at the actual competition instead of a number.

Step 7: target long-tail keywords first

New sites should build on long-tail keywords — specific phrases of four or more words, such as 'how to back up a WordPress site without a plugin' rather than 'WordPress backup'. Long-tail keywords have three advantages: less competition, clearer intent (you know exactly what the searcher wants), and faster rankings. A new site can rank for a good long-tail phrase in weeks; ranking for a short head term can take years.

The trade-off is volume. A long-tail keyword might bring 50 visitors a month instead of 5,000. But fifty visitors who found exactly what they searched for are worth more than five thousand who bounced. And long-tail articles compound: twenty of them, each bringing a trickle of targeted traffic, add up to a real audience — and they build the topical authority that eventually lets you compete for the bigger terms.

Step 8: build a keyword map

Take your list and group it. Each group is one article: a primary keyword, three to five secondary phrases that belong in the same piece, and a note on intent — is the searcher learning, comparing, or ready to buy? A map for a hosting site might look like: 'how to back up a WordPress site' (primary), with 'WordPress backup plugin free', 'restore WordPress from backup', and 'how often to back up WordPress' as secondary phrases in the same article. One thorough article per group. Ten thin articles targeting ten near-identical keywords compete with each other; one thorough article wins.

Keep the map in a spreadsheet with four columns: keyword, intent, difficulty (your judgement from step 6), and status. Work through it in order of easiest wins first. When you finish an article, mark it done and move on — the map is what stops you staring at a blank page every Monday.

A worked example: from seed to article in one sitting

Here is the whole method on one seed, so you can see how fast it runs. Seed: 'start a blog'. Autocomplete gives 'how to start a blog for free', 'how to start a blog and make money', 'how to start a blog on WordPress'. People Also Ask adds 'how much does it cost to start a blog?' and 'can you start a blog with no experience?' Related searches add 'blogging for beginners' and 'best blogging platform for beginners'.

Search Console shows your site already ranks 14th for 'how to start a blog for free' — a page-two opportunity. Trends shows 'start a blog' spiking every January, so a January publish date would catch the wave. Keyword Planner puts the head term in the thousands of searches a month and the long-tail variants in the low hundreds. The difficulty check shows the top ten for 'how to start a blog and make money' is dominated by major publications, but 'how to start a blog for free' has two small blogs and a dated 2023 guide in the top five — beatable.

The map entry writes itself: primary keyword 'how to start a blog for free', secondary phrases 'how much does it cost to start a blog' and 'start a blog with no experience', intent informational, difficulty medium-low. One article, one sitting of research, and you know exactly what to write and why it can rank. Run this loop for each of your ten seeds and you have a quarter of content planned.

When to pay for a tool

Free methods break down at scale. If you need bulk data for hundreds of keywords, competitor gap analysis, or rank tracking across a large site, a paid tool earns its price. Mangools is the budget option most small sites reach for first — its KWFinder tool is built around exactly this workflow, and annual billing cuts the price by about a third. But do not buy a tool to avoid doing the thinking. The free method above finds the keywords; the tool just finds them faster.

Disclosure: the Mangools link above is an affiliate link — if you subscribe through it, SaaSNivo may earn a commission at no extra cost to you.

The bottom line

Free keyword research is not a lesser version of the real thing. Google autocomplete shows real phrasing, Search Console shows your real rankings, Trends shows real timing, and a two-minute look at the top ten results tells you whether you can compete. The paid tools add speed and scale, not a different answer. Start with the free method, build your map, and write the articles — that is the part no tool can do for you.

Once you have your keywords, the next question is how long each article should be. Our guide to how long a blog post should be in 2026 answers that with data, and if your pages need structured data to stand out, read what schema markup is and whether you need it.