How to Do Keyword Research Without Any Tools (Free Methods That Still Work)

Keyword research without Ahrefs or Semrush is not a compromise — it's a different workflow. For most sites with under 50,000 monthly visits, the free methods produce a more than sufficient keyword list. The paid tools add volume estimates and competitor gap analysis; they don't fundamentally change what keywords you should be targeting.

Here are the 7 free methods that produce a real, usable keyword list, in the order you should use them.

Method 1 — Google Autocomplete

Start here. Every suggestion Google shows as you type is a keyword with real search volume behind it. Google's autocomplete algorithm surfaces queries based on actual search patterns from real users — not estimates, not guesses.

Type your seed keyword and pause before hitting enter. Write down every suggestion. Then add a space after the keyword and type each letter of the alphabet, one at a time, writing down what appears. This is the alphabet soup method and it produces 26 batches of keyword variations per seed.

The long-tail keyword research with Google Autocomplete guide covers this method in detail, including free tools that automate the 26-letter process.

Relative power: Very high for long-tail discovery. Limited for understanding volume or competition.

Method 2 — People Also Ask

After you search your seed keyword, Google shows a "People Also Ask" box with 3–4 related questions. Click any question and the box expands to reveal more. Click one of the new questions and more appear. The box is essentially an infinite graph of related queries.

Mine it by recording every question that appears across 2–3 levels of expansion for each seed keyword. These questions are keywords — often long-tail, often zero reported volume in tools, and often exactly what someone in your audience is asking.

Each question you harvest is also a potential H2 or FAQ schema entry for your content. The PAA keyword research method covers the full harvesting workflow.

Relative power: High for question-format content discovery. Excellent for identifying FAQ and featured-snippet opportunities.

Method 3 — Related Searches

Scroll to the bottom of any Google SERP and you'll find 6–8 "Related searches." These are variations and adjacent topics that Google has associated with your query. Search one of them and repeat — each SERP gives you 8 more.

Related searches tend to surface adjacent topics rather than direct variations, which makes them useful for expanding your topic cluster beyond the obvious keywords.

Relative power: Medium. Good for broadening scope, not deep long-tail discovery.

Method 4 — Google Search Console

If your site is connected to GSC, you already have the most valuable keyword data available: real impressions and clicks for queries where your pages appeared in search results. This is site-specific and non-transferable — you can't see competitors' data — but for understanding what to optimize and expand on, it's better than any paid tool.

The full GSC-only keyword research workflow covers the five-step process: pulling opportunity queries, finding declining rankings, surfacing content gaps, and working around GSC's data limitations.

Relative power: Very high for existing-page optimization. No value for sites with no existing traffic.

Method 5 — Reddit and Quora

Search site:reddit.com <your topic> and site:quora.com <your topic> in Google. Read thread titles and question headlines. The phrasing real users choose to describe their problems — not marketing language, not industry jargon — is often the exact keyword phrase they also use in Google Search.

A Reddit thread titled "How do I stop my freelance clients from ghosting me after I send the invoice" is a keyword. "Freelance invoice ghosting solutions" is not a keyword anyone types. The first phrasing is how the problem lives in the searcher's head.

Copy the most common question phrasings, run them through Google Autocomplete to find the exact searchable variation, and add them to your list.

Relative power: High for intent discovery and identifying the way your audience talks about problems. Medium for generating raw keyword volume.

Method 6 — Competitor Sitemap Analysis

Your competitors' content strategies are public. Go to yourcompetitor.com/sitemap.xml (or try sitemap_index.xml). Most CMS platforms generate sitemaps automatically. What you'll see is every URL the competitor has published.

Sort the URLs by folder structure and look for topic clusters you haven't addressed. A competitor with 20 pages on a sub-topic you've written one page about is telling you something about where the keyword opportunity sits.

This method doesn't give you volume data, but combined with Autocomplete (step 1), it's a strong directional signal: if a competitor has invested significant content in a topic, the topic has real search demand.

Relative power: High for strategic gap analysis. Requires manual interpretation.

Method 7 — ChatGPT or other LLMs for clustering

Once you have a raw keyword list from methods 1–6, organizing it into clusters is tedious. LLMs are surprisingly good at this. Paste your list (or a representative sample) and ask:

"Group these keywords by topic and search intent. Label each group with the primary topic and whether the intent is informational, commercial, or transactional."

The output won't be perfect — review it and adjust — but it turns a 200-item raw list into a structured cluster map in minutes instead of hours. This is not keyword research; it's keyword organization. The research itself has to come from sources that know what people actually search for.

Relative power: High for clustering and organization. Zero for generating keyword ideas (LLMs hallucinate volume and search trends).

The full free workflow, in order

1. Write 10–20 seed terms — your core topics, product features, audience pain points.
2. Run Autocomplete on each seed — alphabet soup method, record all suggestions.
3. Mine PAA on your top 5 seeds — 2–3 levels deep, record every question.
4. Check related searches — bottom of the SERP for each seed, follow 1–2 branches.
5. Pull GSC data (if you have it) — impressions by query, identify opportunity clusters.
6. Scan Reddit/Quora — for 3–5 of your highest-priority topics, look for real-user phrasing.
7. Check 2–3 competitor sitemaps — identify topic gaps.
8. Cluster with an LLM — organize raw list by topic and intent.
9. Map to pages — assign each cluster to one existing page or one planned page.

The result is a keyword map built from real-world search signals, zero dollars spent. For most solo bloggers and small businesses, this produces enough material for 6–12 months of content.

When to graduate to paid tools

Free methods have real limits. They don't give you volume estimates, keyword difficulty scores, or competitor ranking data for domains you don't own. You'll want a paid tool when:

• You're producing content at scale (20+ new pages per quarter) and need to prioritize efficiently
• You're in a competitive niche where knowing exact difficulty scores before investing in content matters
• You want to do systematic competitor gap analysis (which keywords rank for them but not for you)
• You're running an agency and need reportable data for clients

Until then, the free stack is more than sufficient.

After building your keyword list and writing content, use RankCrab's free keyword density analyzer to verify each page is covering its target cluster at the right frequency. It takes 30 seconds and catches the most common over-optimization and under-optimization mistakes.

The complete keyword research framework — including the intent-mapping and page structure steps — is covered in the keyword research hub.