Keyword research with free tools: find what people search

What is keyword research, really?

Keyword research is discovering the exact words and questions your audience types into search engines, then choosing which to target based on relevance, demand and winnability. It's market research disguised as SEO — the keywords reveal what people want, in their own language, which shapes not just rankings but products, content and messaging.

Strip away the tooling and keyword research answers three questions: what do people search (the keywords), why (the intent behind them), and can I win (whether ranking is realistic for my site). Everything else — volumes, difficulty scores, SERP features — serves those three.

The vocabulary: seed keywords (your starting topics), long-tail keywords (longer, specific phrases like 'compress image to 100kb for nadra' — lower volume, higher intent, easier to win), search intent (informational / navigational / commercial / transactional — *why* they search), SERP (the results page you're trying to appear on), and keyword difficulty (how hard ranking looks — treat paid scores as rough guesses, not physics).

Why free tools suffice for most sites: the expensive platforms mainly add convenience and historical data. The core insights — what people ask, how they phrase it, who's currently ranking — are all available free. This site's own strategy runs on free research: high-search, low-competition topics found with the methods below, which is exactly how 651 tools got their audiences.

Mindset shift: you're not 'finding keywords to stuff into pages'. You're learning your audience's language so your pages answer their actual questions. The keyword suggestion tool on this site is built for exactly this — real query suggestions, free, no account.

How do you do keyword research with free tools?

Start with seed topics from your business, expand with Google autocomplete/People Also Ask/related searches, the keyword suggestion tool and Search Console queries, then filter by relevance, intent match and realistic competition. The whole workflow is free — method beats software at this stage.

The workflow, start to finish:

• 1. Seeds (15 min). List your topics as customers describe them — not your jargon. A photographer: 'wedding photography lahore', 'couple shoot prices'. Write 10–20 seeds.
• 2. Expand with Google itself. Type each seed and harvest: autocomplete suggestions, 'People also ask' boxes, 'Related searches' at the bottom, and the refinements Google shows. This is Google telling you what people actually search — free and authoritative.
• 3. Expand with tools. Run seeds through the keyword suggestion tool for structured suggestions; check Google Trends (free) for seasonality and rising vs dying topics; mine Search Console queries if the site exists (real impressions you already get).
• 4. Harvest questions. 'People also ask', AlsoAsked-style question mining, Reddit/Quora threads in your niche — questions reveal intent and become FAQ sections and article headings.
• 5. Note the SERP. For each candidate keyword, look at what's ranking: big brands? forums? thin content? The SERP tells you what Google thinks the intent is — and whether you can compete (next section).

Organize into a simple sheet: keyword | intent | who's ranking | my angle. A hundred researched keywords in a spreadsheet beats a thousand exported rows you never read. Research you don't act on is procrastination with extra steps.

YouTube and Amazon deserve mention: YouTube autocomplete + the YouTube tag generator for video topics; Amazon autocomplete for product keywords (shoppers' language differs from Googlers'). If your audience lives on those platforms, research where they search.

How do you judge keyword competition honestly?

Judge competition by examining the actual SERP: the authority of ranking domains, the quality and depth of ranking content, and whether the intent is truly served. A keyword is winnable when ranking pages are forums, thin content, outdated articles or off-topic — regardless of what a difficulty score claims. The SERP never lies; scores often do.

Paid 'difficulty scores' are estimates from third-party link data — useful directionally, misleading precisely. The honest method is reading the SERP like a scout:

• Who ranks? Big brands with massive authority (hard) vs forums, Quora, Pinterest, thin affiliates (opportunity) vs nobody relevant (wide open — but check there's actually demand).
• How good is the content? If the top 3 are comprehensive, current and expert — respect that. If they're 400-word listicles from 2019, outdated specs, or forum threads — a genuinely better page can win.
• Does anything serve the intent well? Sometimes nobody does — the results are adjacent but not answering. That's the best opportunity: be the first to truly answer.
• SERP features. Featured snippets, video carousels, map packs — they change the click math. A keyword 'won' at position 4 below a snippet and videos gets little traffic; factor features into value, not just rankability.
• Freshness signals. Forums and Q&A ranking means Google wants fresh perspectives — new quality content is welcomed. Entrenched 10-year-old authorities ranking means a longer fight.

The realistic matrix for a new/small site: target long-tail, specific, low-competition keywords first ('compress image to 20kb for online admission'), win them, build topical authority, then climb to broader terms. New sites targeting 'image compressor' head-on fight established giants; the same site wins dozens of long-tails that compound into authority for the bigger fight later.

Document your judgment per keyword: 'winnable — top results are forums + thin content; my angle: complete guide with free tool'. This turns research into a battle plan instead of a data dump.

How do you pick the right keywords to target?

Pick keywords scoring well on all three: relevance (would this visitor become your customer/reader?), intent match (can your page format satisfy it?), and winnability (can you realistically outrank the SERP?). Prioritize commercial-intent keywords for revenue pages and informational keywords for authority content — and map one primary keyword per page to avoid cannibalization.

Score candidates on the three axes:

• Relevance. 'Free' traffic that never converts is vanity. A keyword adjacent to your offering ('how to compress images' → image compressor tool) beats a bigger unrelated one. Ask: if I rank #1, does it help the business?
• Intent match. Informational ('how to...') → guides and tools. Commercial ('best X for Y') → comparisons. Transactional ('buy', 'near me', brand + product) → product/store pages. Mismatched format never ranks — Google knows what searchers want.
• Winnability. From the SERP analysis above — be honest about your site's current authority. New site? Long-tails only. Established? Broader terms in your topical wheelhouse.

Map one primary keyword per page. Two pages targeting the same keyword split your authority and confuse Google (cannibalization) — consolidate into the stronger page. Your sheet should show a clean mapping: each target keyword → exactly one URL (existing or planned).

Balance the portfolio: 60% winnable long-tails (traffic now, authority building), 30% medium-competition in your expertise zone (traffic soon), 10% aspirational broad terms (the long game). Review quarterly: promote winners, rewrite strugglers, retire losers. Keyword research isn't a document — it's a pipeline.

Pakistan/India note: research in the audience's actual language mix — Roman Urdu/Hinglish queries ('image compress karne ka tarika'), transliterated terms, and English-with-local-context ('gold rate today lahore') are real high-volume queries that English-only research misses. The autocomplete harvest catches these naturally if you seed with local phrasing.

What do you do after choosing keywords?

After choosing: create or optimize one page per keyword following the on-page checklist (intent-matching format, complete coverage, question headings with direct answers), interlink topically, then track rankings and clicks in Search Console and iterate. Keywords are hypotheses — rankings and traffic data confirm or kill them, usually within 4–12 weeks.

Execution per target keyword:

• Build the page the SERP wants. Format matched to intent, coverage deeper than current winners, genuine added value (your data, examples, tools). Follow the on-page SEO checklist — titles, headings, depth, internal links, images, schema.
• Answer the questions. Use harvested 'People also ask' questions as H2s with 40–60 word direct answers — the AEO pattern that wins snippets and voice answers.
• Interlink. New page gets links from relevant existing pages; it links out to related pages. No orphans.
• Titles/descriptions sized right via the meta tag generator — CTR is part of the game.
• Track. Search Console: impressions, clicks, average position per query. Give it 4–12 weeks.

Then iterate on data, not hope: rising but not top-3? deepen content, add the subtopics competitors cover, improve internal links pointing to it. Impressions but no clicks? rewrite title/description for CTR. No impressions? check indexing, then reassess winnability — maybe the keyword was harder than judged. Ranking but bouncing? intent mismatch — rebuild the page to the intent.

Content refresh cadence: revisit target pages every 6–12 months — update data, extend coverage, refresh examples and dates. Rankings decay as competitors improve; maintenance is the moat. (This site's rule: honest updated dates from real changes, never fake freshness.)

Scale sensibly: one excellent page per week beats five thin ones. The compounding math of SEO rewards consistency over bursts — twelve strong pages a year, each targeting researched keywords, builds more than a hundred unresearched ones.

What are the biggest keyword research mistakes?

The biggest mistakes: chasing volume while ignoring intent, trusting difficulty scores over SERP reading, targeting keywords your page format can't satisfy, cannibalizing with multiple pages per keyword, researching once and never revisiting, and ignoring how your specific audience phrases queries (especially local language mixes). Research is only as good as its honesty.

Volume blindness. '10,000 searches!' for a keyword whose intent you can't serve (or whose visitors never convert) is worthless. A 200-search keyword with perfect commercial intent beats it every time. Volume × relevance × winnability — all three, always.

Score worship. Difficulty 12/100 that ranks three DR-90 giants with perfect content is not easy; difficulty 45/100 with forums and thin content on top is winnable. Read SERPs, not scores.

Format mismatch. Targeting 'buy X' with a blog post, or 'how to X' with a product page — Google has shown you the winning format; arguing with the SERP is arguing with your customers.

Cannibalization. Five blog posts all 'about' the same keyword, none ranking — consolidate into one definitive page and redirect the rest. Audit with a site: search or Search Console overlap.

One-and-done research. Language shifts, competitors improve, new questions emerge. Refresh research yearly; your keyword sheet is a living document.

Ignoring local phrasing. Pakistani and Indian audiences search in mixes — English, Roman Urdu, transliterated terms, city-modified queries. Seed and harvest in their language and you'll find low-competition keywords the English-only crowd never sees — this is a genuine structural advantage for local sites.

Frequently asked questions