Keyword research for a product nobody is searching for yet

You built something genuinely new, you type it into a keyword tool, and the tool says zero. This is the single most common way keyword research goes wrong for a young product, and the usual advice — "target long-tail keywords" — is not advice, it is a shrug.

Here is what actually works.

Stop searching for your product

Nobody searches for a category that does not exist yet. In 2009 nobody searched for "async standup tool"; they searched for "how to run a standup with a remote team". The first phrase is your product. The second is your customer's Tuesday morning.

So run your research on the problem, not the solution. Concretely, seed with:

• what the person was doing immediately before they needed you
• the manual workaround they use today, including the spreadsheet
• the incumbent they are frustrated by, by name
• the outcome they would describe to their boss

For a hypothetical tool that turns customer calls into changelog entries, those seeds are "how to write release notes", "changelog examples", "notion release notes template" and "customer feedback to roadmap". Not one of them contains the product.

Read intent before volume

Sort your results by intent, not by volume, at least on the first pass.

A phrase like changelog examples is informational — someone is orienting. Release notes tool is commercial — someone is shopping. Notion release notes template is transactional dressed as informational: they want a thing they can use in the next ten minutes, and if that thing is yours, you have them.

The transactional-in-disguise phrases are where a small product wins, because the incumbents write listicles for them and you can write the actual template.

The volume floor is lower than you think

A phrase with 90 searches a month sounds like nothing. If it is unmistakably your buyer, 90 is fine. Ten per cent of them clicking is nine visits, two per cent of those converting is a customer every five months from one page you wrote once. Do that thirty times.

The failure mode is not low volume. It is low volume plus no purchase intent — writing about "what is a changelog" and getting students.

What this costs, concretely

Four seed phrases at 250 rows is four searches: 84 credits, most of the 100-credit free signup grant. That is a full afternoon's research budget, and it is the actual arithmetic rather than a hypothetical.

Then save the twenty phrases that survive the read-through — saving is free — tag them by the page you would write, and let the list sit for a week. The ones that still look good after a week are the ones to write.

Then check whether you were right

Six weeks after the pages are live, track those twenty phrases weekly. Twenty keywords on one device is twenty keyword slots — well inside the free plan's 25, and it costs no credits at all.

What you are looking for is not position one. It is movement: page four to page two means the topic accepts you and the next page on the same theme will land faster. Flat at position 80 for two months means the topic does not, and you should stop writing about it and go find the next problem your customer has on a Tuesday morning.