Keyword research: finding what people actually type

Keyword research is the work of finding out what people type when they have the problem you solve. It is less about discovering clever terms than about accepting plain ones: the phrases that bring customers are usually duller and more specific than anything a marketing session produces, and they are written in the customer's vocabulary rather than yours.

What keyword research has to establish

• Intent — whether somebody typing this wants to learn, to compare, or to buy. A page that answers the wrong one of those will rank and convert nothing.
• Competition, realistically: whether the current results are pages you could plausibly beat, not a score from a tool.
• Volume, which matters least and is treated as mattering most.
• Whether you can genuinely answer it better than what already ranks, which is the only durable advantage.
• Whether the term describes something you actually sell, since traffic for everything else costs money and returns nothing.

Why volume is the least useful number

Because it is an average across a year, rounded into bands, and shared with everybody chasing the same list. A term with a thousand searches a month where the first page is occupied by large competitors returns nothing; a term with forty searches where somebody is describing their exact problem can bring a customer a week. Small companies win on specificity, which means deliberately choosing lower-volume phrases that match what you do precisely — and the tooling pushes the opposite way, because big numbers look like opportunity.

Look at the results page before judging any term. If the first page is dominated by comparison sites and marketplaces, the search engine has decided what kind of page belongs there, and an ordinary article will not displace it regardless of quality. That single check saves more wasted writing than any metric.

Where the winnable terms hide

1. In the words customers use in enquiries and support tickets, which rarely match the industry term.
2. In questions: anything starting how, what, when, or whether somebody needs something.
3. In the specific over the general — a term with three words beats the same topic in one.
4. In the things you get asked on the phone, which nobody writes down and everybody searches.
5. In your own existing pages, which already rank for terms you never targeted and can be improved cheaply.
6. Not in competitor keyword exports, which show what they chose rather than what worked.

Turning a list into pages

One intent per page, and a page that answers the question completely rather than ten pages circling it. The commonest failure after good research is splitting a topic across several thin pages targeting near-identical phrases, which makes them compete with each other and ranks none of them. Group the variants that mean the same thing, write one page that would satisfy somebody asking any of them, and move on to the next genuinely different question.

Ettex Web keeps the page and the thing it was written to answer together, which makes the later check possible: did this page bring anybody, and for what. Whether it did belongs with the conversion tracking rather than with rankings, since a page that ranks and converts nobody is still the wrong page.