For an SMB, a useful keyword list runs to between fifteen and forty phrases, not five hundred. You keep them on three criteria, in this order: intent — can the person typing this become a customer — real volume in your region, and difficulty. A phrase with forty monthly searches and buying intent beats one with three thousand purely informational searches. That is the first thing to accept, and ignoring it wastes more effort than anything else on the list.
This piece sets out the method we use to build that list, including for our own content. It complements our piece on content marketing, which deals with the why; this one deals with choosing the subjects.
Intent before volume
Three people type three very different things. Someone asking what engineered flooring is wants to understand. Someone comparing engineered flooring against hardwood is weighing options. Someone searching for a flooring installer in their town wants to call somebody. The third is worth ten times the first two to a service SMB, even if it generates ten times fewer searches.
That does not make the first two useless: they catch people who will buy in six months, and they build the credibility that tips the decision later. But they come second. An SMB starting with informational subjects waits two years to see a customer, and usually concludes that search does not work.
Where to find the phrases, free
Four sources are enough for a first list. The first, and by far the best: your own Search Console. It tells you which queries you already appear on, at what position, and how many people click. It is the only source talking about your site rather than the market in general — and the one most businesses never open.
The second: your customers' own words. Re-read twenty quote request emails and note the exact phrasing. It almost never matches your trade vocabulary, and that gap is precisely why people do not find you. The third: Google's autocomplete suggestions and the related questions block, which read demand directly. The fourth: the pages of your two or three real competitors, the ones you meet when quoting.
The word you use for your trade is almost never the word a customer types to find it.
Reading the numbers without getting them wrong
Three indicators show up in every tool, and two of them are regularly misread. Volume is a monthly average, often national: a phrase showing 2,000 searches across Canada may be worth 60 in your region. That is the most frequent mistake, and the most expensive one to build a plan on.
Difficulty, scored from 0 to 100, estimates the effort needed to reach the first results. Under 20, an SMB can get there with a good page. Above 50, it takes time and links. The third number, advertising competition, does not measure organic difficulty at all: it measures how many advertisers are bidding. High advertising competition is in fact a good signal — it proves the phrase pays well enough for people to buy it.
The best ratio always sits at the intersection: decent volume, low difficulty, commercial intent. Building our own editorial plan, the most profitable phrase we measured showed 6,600 monthly searches at a difficulty of 7 — a case with nothing to weigh up. Situations like that exist in nearly every sector, but you do not find them without opening the tools.
The opposite trap: phrases nobody types
There is a difference between having something to say about a subject and that subject being searched for. We tested it on our own article ideas: three subjects out of four where we had real expertise showed no measurable volume. The phrases simply did not exist — nobody types the name of a niche trade followed by the word marketing.
Those subjects keep their value as proof of expertise, in a case study or on social platforms. They have none as a traffic lever. Checking volume before writing takes ten minutes and saves weeks of writing for no reader.
One intent, one page
Once the list is built, it has to be distributed. The rule is simple: one page per intent, never two pages on the same one. Two pages aiming at the same phrase compete with each other, and Google ends up ranking neither properly. Close phrases — bathroom renovation and redoing a bathroom — go on the same page; different intents go on different pages.
Second rule: commercially intended phrases belong on service pages, not on articles. Someone looking for a supplier wants a page that sells, not a thousand words. Articles serve informational intent and point back to the service pages. That is what separates a blog that pays from a blog that merely exists.
What generated answers change
A share of purely informational queries now gets answered directly in the results page or inside an assistant, with nobody clicking. That is one more argument for favouring commercially intended phrases, where the visitor still needs to go and look at a supplier. And for writing in a way that gets picked up by those systems rather than bypassed — that subject is handled separately.
If you do not have a list yet, a decent starting audit should contain one: it is one of the five areas described in what a free SEO audit should contain. Once the list exists, turning it into a workable writing plan is the subject of our method for building a content strategy. Our full offer is on the SEO and GEO page, and the budget orders of magnitude in what web marketing costs in Quebec.
And if you would rather we looked together at what your customers actually type to find you, book the call: we open your Search Console during it, which settles the question in ten minutes.
Defining your ideal customer: the eight-line sheet — knowing who you are talking to before choosing the words.