Keyword Research Without Expensive Tools

A keyword tool subscription costs somewhere between the price of a phone contract and the price of a junior salary. Before you sign up for one, it is worth knowing what you are buying. You are buying convenience, competitor data, and tracking. You are not buying the search volumes — those are modelled estimates, and every tool models them differently, which is why three tools will give you three numbers for the same phrase.

None of the free sources give you a tidy spreadsheet. All of them give you something the paid tools cannot: what people typed to reach your site, and what the results page actually looks like for the term you care about.

Start with the searches you already appear for

Search Console is the only source of first-party query data that exists, and it is free. Open the Performance report, set the date range to the last three months, and go to the Queries tab.

Three filters make it useful rather than overwhelming:

Average position between 8 and 20. These are the phrases you nearly rank for. Improving a page from position eleven to position five is far less work than creating something new from nothing, and the demand is already proven because you have the impressions to show it.

High impressions, low clicks. Either the page is not appealing enough in the results — a title and description problem — or the query means something different from what your page offers. Both are worth knowing, and they need different fixes.

Queries with no matching page. If a phrase is generating impressions but the URL Google shows for it is your homepage or a loosely related post, that is a page waiting to be written.

Export the query list. It is the closest thing to ground truth you will get.

Use the results page as the research tool

Search the phrase yourself, in an incognito window, and read what comes back. This tells you things no volume figure can.

Look at the type of result. If a commercial phrase returns ten informational guides, Google has decided that searchers want to learn rather than buy, and a service page will struggle there. If it returns a map pack and three directories, the win is a Google Business Profile rather than a blog post.

Look at who ranks. Ten national brands with dedicated content teams tells you the phrase is expensive to compete for, regardless of what a difficulty score says. Two competitors of roughly your size and a couple of thin directory listings tells you the opposite.

Then mine the page itself. The autocomplete suggestions as you type are real queries. “People also ask” gives you the questions attached to the topic, and each one expands into more. The related searches at the foot of the page give you phrasing variants you would not have guessed. None of these come with volumes, which is fine — they come with confirmation that people search this way.

Keyword Planner, used carefully

Google’s Keyword Planner is free with an Ads account. You do not have to run a campaign to open it, though accounts that have never spent are shown broad ranges rather than tighter figures.

Two things to know before you trust the output.

It groups near-identical terms together, so singular and plural variants often report the same number even when they behave differently in the results. And the ranges are wide by design. Treat them as an order of magnitude — is this phrase searched dozens of times a month or thousands? — rather than a forecast.

What Planner is genuinely good at is discovery. Feed it a competitor’s URL and it will return the terms it associates with that page. Feed it your own service names and it will return the phrasing variants you had not considered.

Google Trends for direction, not volume

Trends does not report how many people search something. It reports relative interest over time, indexed to itself. That makes it useless for sizing a market and genuinely useful for two other questions: is this term growing or fading, and is it seasonal? A term that spikes every January is worth publishing for in November, not in January.

The comparison feature is the better half of the tool. Put two phrasings side by side and you will see which one your customers actually use.

Ask the people who answer the phone

The most underused source in any business is the language customers already use. Read the last fifty enquiry emails. Listen to how people describe their problem on the phone. Check the internal search box on your own site if you have one — those are queries from people who are already interested and could not find what they wanted.

Customers rarely use industry vocabulary. They describe a symptom, not a service. That gap is where the least contested phrases live.

Judging a keyword without a difficulty score

Difficulty scores are proprietary compressions of link data. You can approximate the judgement without one by asking four questions:

1. Does the results page match what I sell? If not, no amount of optimisation fixes it.
2. Is there anything on page one I could plainly do better? Thin pages, out-of-date pages and directory listings are all beatable. A well-maintained page from a specialist is not.
3. Would someone searching this be ready to enquire? A phrase with modest demand and clear commercial intent beats a broad one that brings browsers.
4. Can I say something true and specific about it? If you cannot, the page will be filler, and filler now carries real risk rather than merely wasting effort.

What paid tools actually add

Being honest about this matters. Three things justify the subscription once you are past the basics.

Competitor visibility. Free sources tell you what you rank for. Paid tools estimate what your competitors rank for and you do not. That gap analysis is difficult to reproduce by hand.

Backlink data. Search Console shows links to your own site only. Seeing who links to a rival, and why, needs an external index.

Rank tracking at scale. Checking twenty phrases manually is tedious. Checking two hundred across several locations is not something you will keep doing.

If you have fewer than thirty pages and one location, you can go a long way before any of that becomes the constraint. Spend the money on the content instead.

Where this usually goes wrong

The common failure is not the absence of tools. It is picking terms by volume alone, publishing one page for each, and never checking whether the results page ever wanted that page in the first place. Research is a judgement about intent and competition. The numbers are an input to it, not a substitute for it.

If you would rather have someone work through your Search Console data and tell you which pages are closest to earning something, that is where an SEO audit starts. Getting the measurement right first makes everything after it easier to judge.