How Do You Do SEO When Nobody Searches for Your Category?

How do you do SEO when nobody searches for your category?

You stop searching for your category and start searching for the problem it solves. Nobody types a term they have never heard. They type the symptom, the workaround they are using now, or the older category they think they are shopping in. Those searches have volume, and they belong to you.

This comes up constantly with new B2B products. The founder has a name for what they built. The keyword tool returns a flat zero. The marketing team concludes that search is not a channel for them and moves the budget to ads.

We think that conclusion is usually wrong, and the reason is that the keyword tool is answering a different question than the one that matters.

Why does a zero volume category happen at all?

Because naming something new is the whole point of category creation, and a new name has no search history by definition. A keyword tool reports the past. A new category has no past. The zero is accurate and also useless for planning.

There is a second cause that is less flattering. Sometimes the category is not new, it is just internal language. The company says workflow orchestration layer and the buyer says I need my spreadsheets to talk to each other. That is not a missing category. That is a positioning problem wearing a keyword costume.

Telling those two apart is the first job. If a buyer would recognise your category name in a conversation, it is real and early. If they would ask you what it means, you have an internal phrase and you should not build a content plan on it.

Should you target the volume that exists nearby?

Yes, and this is where most of the traffic will come from for a long time. Map the searches around your product rather than for it. That means the problem people search when they have it, the tools they are trying to replace, the manual process they are doing today, and the job title that owns the pain.

Those four buckets almost always have real volume, because they describe things people already know they have. A buyer who searches how to stop double entry between two systems is your buyer, even if they have never heard your product name and never will before they land on the page.

The trap is writing those pages in your language instead of theirs. If the page opens with your category name, the reader bounces. Open with their problem in their words, then introduce the name later, once the page has earned the right to teach them something. Our notes on search intent go deeper on matching the page to what the searcher actually wanted.

What is a problem first keyword map?

It is a keyword plan built from the buyer's situation rather than your product taxonomy. You start with five or six moments where the problem becomes visible to them. Then you list the phrases a person in that moment would actually type, and only then do you check volume.

Checking volume last matters. When you check it first, you unconsciously bend the plan toward whatever the tool rewards, and you end up with a blog full of generic posts that rank for terms your buyers never search. When you check it last, you keep the shape of the buyer's journey and accept that some pages exist for a small, precise audience.

We usually end up with a small number of high volume problem pages, a larger number of specific low volume pages, and one page that defines the category properly for the people who arrive already curious. That third one rarely gets traffic and often gets cited, which is a fair trade.

Do zero volume keywords ever pay off?

They pay off in aggregate and in quality, not individually. Ahrefs looked at this directly in a study published on 21 October 2022. Pages ranking in the top five for keywords with a search volume of 10 or lower drove an average of 11.3 impressions a month. Fewer than 1 percent of those keywords drove more than 100 impressions.

Ahrefs was careful to note the sample was small and might not represent every industry, and we would repeat that caution. The useful part is the shape rather than the exact number. A zero volume page is not a traffic play. It is a presence play.

There is a reason to believe the tools undercount as well. Google has said on its own Search blog, in a post published on 25 April 2017, that 15 percent of the searches it sees every day are new. If a sixth of daily queries have never been typed before, no keyword database can hold them. Your category term is sitting in that gap.

How does AI search change this problem?

It helps more than it hurts, because answer engines invite full sentences. Somebody describing a problem to ChatGPT writes three lines of context that they would never type into a search box. That is exactly the kind of query a new category can win, since the engine is matching meaning rather than an exact phrase.

The audience is large enough to take seriously. Similarweb's AI search report of 29 July 2026 found AI platforms drew an average of 9.5 billion web visits a month worldwide between June 2025 and May 2026, roughly 70 percent growth year over year.

There is also more room in an AI answer than on a results page. Semrush's 2026 AI Visibility Index, released on 26 June 2026 after analysing 126 million United States AI search prompts, found ChatGPT cites an average of 15 sources per response. Gemini cites an average of 3. Fifteen slots is a genuine opportunity for a company nobody has heard of.

Should you name the category or borrow one?

Borrow first, name later. Early on, ride an existing category that your buyer already shops in, and be the option in it that solves the problem differently. Once you have customers who describe you in their own words, take the words they use and make that the name.

Naming first is the expensive mistake. You spend a year teaching a term while your competitors collect the demand that already exists under the old term. Plenty of good products have died on that hill with a beautiful brand book.

The compromise we like is a clear definition page for the new term, linked from pages that rank for the old one. The old term brings people in. The definition page gives the idea somewhere to live, and gives answer engines something clean to quote. Our piece on B2B SaaS positioning covers how to pick which category to fight in.

Where should this content actually live?

On your own domain, in a structure a crawler can follow without guessing. Problem pages near the top, deeper specific pages linked from them, and the definition page linked from everything that touches the idea. Flat and obvious beats clever.

Ranking still helps even in AI answers, though less than it did. Ahrefs published an update on 2 March 2026 covering 863,000 keyword search results pages and 4 million AI Overview URLs, and found that 38 percent of cited pages also rank in the top 10, down from roughly 76 percent in its July 2025 study. So you can be cited without ranking, but ranking is still the strongest single signal you control.

Keep the pages fast and readable without JavaScript. A page that needs a script to show its own text is a page an engine may skip. Our guide to topical authority covers how the set of pages works together.

How do you measure progress when volume is near zero?

You change the metric before you start, or you will kill the project in month four. Sessions is the wrong headline number here. Better ones are impressions on problem terms, branded search growth, and how often your definition page shows up as a source in AI answers.

Branded search is the one we watch hardest. If people start typing your company name after reading your problem pages, the positioning is landing. That number is visible in Google Search Console and it moves slowly and honestly.

Set the review point at six months, not six weeks. Category work is slow by nature, and judging it on a monthly traffic chart guarantees you abandon it right before the compounding starts.

What would we do in the first month?

Three things. Interview five customers and write down the exact words they used for the problem before they found you. Build the problem first keyword map from those words. Then publish the two highest volume problem pages and one clear definition page, properly sourced.

That is enough to test whether the demand is there in some form. If those pages get impressions, the channel works and you scale it. If they get nothing after a fair run, the issue is probably positioning rather than search, and no amount of publishing will fix it.

If you are staring at a keyword tool full of zeroes and trying to work out whether search is worth the effort, we are happy to talk it through. You can find us at phoenix.studio.