Keyword research without the enterprise tool bill: a working method

AUGUST 7, 2026 · 7 MIN READ · SEO / KEYWORD RESEARCH / PLAYBOOK

You can do genuinely good keyword research with free sources: customer language (support tickets, sales calls, community threads), Google’s own suggestions (autocomplete, People Also Ask, related searches), and Search Console queries you already almost-rank for. The expensive tools mostly add volume estimates and speed — and the volume numbers are estimates from every vendor, including the expensive ones.

Start with customer language, not a tool

The classic failure mode of tool-first research is a spreadsheet of five hundred keywords in the vendor’s vocabulary, none of them in your customer’s. Start from the words real people use when they have the problem you solve:

• Support tickets and sales calls. How do people describe the problem before they know your terminology? Those phrasings are queries.
• Communities.Reddit, forums, Discord servers, Facebook groups in your niche. The titles of recurring questions are keyword research that someone else did for you, in the searcher’s own words.
• Your own inbox. The questions prospects ask before buying are the questions other prospects are typing into Google right now.

Twenty phrases from these sources beat two hundred from a suggestion engine, because every one is attached to a human you know exists.

Expand with Google’s free hints

Google tells you what people search for; it just does not hand it over in a CSV.

• Autocomplete.Type a seed phrase and pause. The suggestions are real queries. Walk the alphabet after your seed (“keyword a”, “keyword b”…) and you have a long-tail expansion for free.
• People Also Ask. Each box is a question Google has decided is closely related — and expanding one loads more. These map naturally to headings and FAQ content.
• Related searches at the bottom of the page show adjacent intents you may not have considered.
• Search Console. This is the underrated one. Filter the Performance report for queries where you average position 8–25. These are keywords Google already believes you are relevant for; you are one better page or one improved section away from real traffic. It is the closest thing to a cheat code in free SEO.

Judge intent by reading the SERP itself

Every keyword tool tries to label intent — informational, commercial, transactional. You do not need the label; you need the evidence, and the evidence is the live results page. Search the keyword and look at what kind of pages rank. That is Google showing you, empirically, what it thinks the searcher wants:

• All blog guides? Google reads it as a learning query — a product page will struggle there.
• Product and category pages? Buying intent — a blog post is the wrong weapon.
• Listicles and comparisons? The searcher is shortlisting — you need a comparison, or to be in someone else’s.
• A wall of ads, maps or shopping units? Note how far down the first organic result actually starts before you fight for it.

If the page type you were planning does not appear in the top ten, the keyword is not wrong — your page is.

Pick battles by looking at who actually ranks

Difficulty scores are a convenience metric, and they routinely mislead in both directions. The honest check is to look at the ten sites currently ranking and ask who they are. If the results are major publishers and category giants with dedicated pages, that is a long war regardless of what any score says. If you see forum threads, thin pages, decade-old posts, or pages answering the query only in passing, that is an opening — Google is ranking them because it has nothing better, and you can be the something better. A small site’s realistic near-term wins are almost always specific, long-tail queries where the current results are visibly mediocre.

Group keywords to pages, one intent per page

The output of research is not a keyword list; it is a page plan. Cluster keywords that mean the same thing — if two queries would be satisfied by the same content, they are one cluster and get one page. Splitting one intent across three thin pages makes them compete with each other; stuffing three intents into one page makes it rank for none. A simple sheet with columns for cluster, primary keyword, supporting variants and target URL is the entire deliverable.

A word on volume numbers

Whatever tool you eventually use: search volumes are modeled estimates, not meter readings, and different tools will confidently give you different numbers for the same keyword. Treat volume as an order-of-magnitude signal — is this a hundreds keyword or a tens-of-thousands keyword — and let customer evidence and the SERP make the actual decisions.

When you want to speed the SERP-reading step up, HawkOrbit’s Keyword Research tool analyzes a keyword against the live results page — who ranks, what kind of pages, how crowded the opening is — which is the same judgment described above, just without the tab fatigue. The method stands on its own either way: customer words in, live SERP as the referee, one page per intent out.