1. Demand

You can be perfectly indexed for a phrase nobody types. That was me: the pages were up, the crawler kept taking them, and nobody searched for that word.

That is why this section stands first. Everything further along the route costs weeks — crawler access, pages, links, dashboards — and none of it pays you back unless somebody is already searching.

One question comes before all of that: what are you building, and how does it differ from what is already on sale? Where a niche starts when everything is taken answers it — look for somebody else’s rough edge, not an empty market — and counts what skipping the step cost me.

Then two more questions live here, and a phrase passes one of them and fails the other easily enough. The words people actually type asks whether anyone searches at all. Whether that audience pays for anything asks whether there is money behind the searching.

People trip over the money. Real volume, an empty ad block, and everybody leaves happy, because they found the free way.

Which market to build for works both questions on real figures. Three niches and four markets, read in Semrush on one day. The niche with the biggest number turned out to be the worst of the three.

The paid tools live here rather than further along, because this is the moment people buy them, and a panel describes a market you have not entered. It describes a site with no traffic far worse, and that is the case I bought one for.

So do both before the writing. I spent a month writing pages around my own word for the thing. Good word. Nobody typed it.