SEO

Keyword Research Without Paid Tools: A Complete Free Workflow

A repeatable keyword research process using only free sources — search suggestions, competitor pages, Search Console and forums — plus how to judge difficulty by eye.

·3 min read·696 words
Search bar with branching keyword nodes

Paid keyword tools are useful, but they are not where good keyword research starts, and they are certainly not required for a small site. The expensive part of keyword research is judgement, and judgement is free. Here is a full workflow that costs nothing.

Step 1: write down what you actually sell or know

Before touching any tool, list the ten to twenty things your site is genuinely about. Not aspirations — things you can write about with authority. This list is your topical core, and every keyword you eventually target should sit inside it. Sites that rank consistently do so because they cover one area deeply, not because they scattered posts across whatever had volume.

Step 2: harvest suggestions from search itself

The search box is the largest free keyword tool in existence.

  • Type your core term and read the autocomplete suggestions. Then type the term followed by each letter of the alphabet for a much longer list.
  • Scroll to "People also ask" and expand a few — each expansion loads more questions.
  • Read "Related searches" at the bottom of the results page.
  • Repeat on YouTube, which surfaces different phrasing because intent there is more instructional.

Collect everything in a spreadsheet without filtering. You are gathering language, not deciding yet.

Step 3: mine the places your customers talk

Reddit threads, niche forums, Facebook groups, product review sections, support tickets and sales call notes are full of the exact phrasing people use before they know the industry jargon. These phrases are frequently lower competition precisely because keyword tools under-report them.

Pay special attention to questions asked repeatedly. A question asked five times in a forum is asked five thousand times in search.

Step 4: read the competition properly

Pick the three or four sites that consistently appear for your core terms. Look at their navigation, their category pages and their most-linked articles. You are not copying topics — you are mapping the shape of the subject and spotting the gaps.

Then, for each keyword you are considering, actually look at the results page:

  • Are the top ten dominated by huge brands, or are there small sites in the mix? Small sites present means the term is winnable.
  • Do the ranking pages genuinely answer the query, or are they thin and outdated? Weak incumbents are an opportunity.
  • What format is winning — a guide, a tool, a comparison, a video? Match the format or you will not compete regardless of quality.
  • Is the intent commercial or informational? Do not write a blog post for a term where every result is a product page.

This visual assessment is more reliable than any difficulty score, because it reflects the actual competitive field rather than a domain-authority average.

Step 5: use Search Console, the most underrated free source

If your site has any history, the performance report shows queries you already receive impressions for. Two goldmines live here:

  • Queries ranking 8–20. You are close. A better title, a stronger section, and a couple of internal links often push these onto page one.
  • Queries with impressions but almost no clicks. Usually a title and description problem, not a ranking problem. Rewriting them is the cheapest traffic you will ever gain.

Step 6: prioritise honestly

Score each candidate on three things: how relevant it is to what you sell, how winnable the results page looks, and how much you actually have to say. Anything scoring low on the third should be dropped, however tempting the volume.

Then group survivors into clusters — one substantial page per cluster, not one thin page per keyword variation. Modern search engines understand synonyms, so ten near-identical pages compete with each other instead of ranking.

Step 7: write, publish, wait, revisit

Give a new page eight to twelve weeks before judging it. Then return to Search Console, see which unexpected queries it picked up, and expand the page to serve them. That feedback loop — publish, observe real queries, expand — is what turns a decent page into a page that owns its topic, and it costs nothing but attention.

#keyword research#seo#free tools

Keep reading