How Do You Market a Product Nobody Is Searching For Yet?

You run the keyword research for a launch and the report comes back almost empty. The head term returns close to nothing. The long tail is misspellings, a competitor’s brand name and a handful of queries that describe a different problem entirely. The pages ranking nearest to what you built belong to a category that does not describe it, and the people reading them want something else.

That is not a research failure. It is the finding, and the plan has to be built on top of it. Category creation marketing starts from an uncomfortable premise: when nobody is searching, no amount of on-page work produces traffic, because there is no traffic to capture. You are not competing for a query. You are trying to make one exist, in other people’s heads, one conversation at a time.

Our position, stated plainly. For a product with no search demand, search is a second-year investment and the first year belongs to problem-first content and other people’s audiences. That is not an argument against building the technical foundations now; they cost little and they compound quietly. It is an argument against measuring the first two quarters in rankings, and against hiring for a channel that has nothing to harvest yet.

What follows is the sequence we would run, the parts of it that reliably disappoint people, and the honest shape of the timeline.

Why category creation marketing has no SEO shortcut

Search is a harvesting channel. It works by standing where demand already flows and being the best answer when it arrives. Every instrument in the discipline assumes the query exists: volume, difficulty, share of voice, the rank tracker itself. Remove the query and the instruments have nothing to read.

Teams reach for three workarounds and all three fail in the same way. They target the category they wish they were in, and collect visitors looking for a different product. They target their own brand name, and reach people who already know them. Or they buy the traffic, which behaves the way paid always behaves once you push it past the people already in market, a pattern we have described in why paid search stops working when you scale the budget.

The alternative is not clever. It is slower, and it is mostly writing things down and talking to people who were not asking. It works because demand for a new category never begins with a product query. It begins with a problem somebody is already solving badly, at some cost, with a tool that was never meant for it.

The three audiences you actually have

Before any plan, separate the people you are talking about. They behave differently enough that one message cannot serve all of them, and the differences decide which channel does the work.

• The workaround crowd. They have the problem, they have named it, and they are searching for a way to survive it: a template, a manual process, a tool from an adjacent category. They are findable today, and they are the cheapest people you will ever reach.
• The unnamed crowd. They have the problem and do not experience it as a problem. It is simply how the job feels. They will never type a query about it, so they have to be met inside something they already read, watch or attend.
• The category tourists. They search the term you eventually want to own, but they mean something else by it. Ranking for them costs money and produces nothing except a bounce rate somebody will misread as a messaging problem.

Most launch plans are written for a fourth audience that does not exist yet: people searching for the product by the category name you intend to invent. Writing for them first is the most reliable way to spend a year producing pages nobody reaches.

Write for the problem, not the product

The first content plan for an unsearched product should barely mention the product. That sounds like advice from a conference stage, so here is the mechanical reason: the only queries that exist describe the symptom, and symptom queries are usually plentiful, unglamorous and cheap to win. Nobody is fighting over them because nobody has worked out what they are worth.

So write the page that answers the question somebody types at eleven at night when the workaround breaks. That is the real entry point to a category that does not exist yet, and it is a page a search engine can already justify ranking, because it answers something a person genuinely asked.

Finding the words people actually use

The vocabulary is already written down, just not in a keyword tool. It is in sales call recordings, support tickets, the subject lines of inbound email, community threads about the workaround, and the review sections of the tools your buyer is currently misusing. Read enough of them and the phrasing repeats itself with unnerving consistency.

The output is a vocabulary list: the nouns your buyer uses for the problem, the verbs they use for the workaround, and the words they never use at all. It feeds messaging and verbal identity at least as much as it feeds the editorial calendar, and it is worth finishing before either one starts.

The page types that earn their place

Not every page is worth writing in the first year. These are, roughly in the order they pay for themselves.

Publish the definition page anyway. It costs an afternoon, it will do nothing measurable for a long time, and it is the page every later citation, comparison and analyst mention eventually points at. Writing it early also forces the team to agree on one sentence, which is harder and more useful than it sounds.

Adjacent queries: demand that already exists

There is always adjacent demand. The work is finding the queries whose searcher shares your buyer’s problem even though the query says nothing about your product.

The first kind is the job. Somebody searching how to do the thing manually is describing your product without knowing it exists. The second is the tool they already pay for: comparison and alternative queries around an adjacent category put you in front of people who have already accepted that this problem is worth spending money on. The third is the consequence, the mess the problem causes further downstream, which is often where the budget holder is looking rather than the user.

Alternative-to pages deserve one rule of their own: write them for somebody who already likes the incumbent. A page that opens by explaining why the tool your reader chose is bad tells that reader they were careless, and readers do not buy from people who have just insulted their judgement. Describe what the tool is genuinely good at, then describe the situation where it stops fitting. The comparison that survives scrutiny is the one that concedes something.

These pages move before category pages do, in both traffic and sales conversations, which is why early reporting should look like reporting SEO progress before the rankings move rather than a rank-tracker screenshot pasted into a deck.

The ordering of what you ask for matters more here than on an ordinary site, because the visitor has no category to file you under yet. We have written about what a landing page should ask for, and the short version applies with force: the first request should cost the visitor almost nothing and tell you almost everything.

Distribution through other people’s audiences

If nobody is searching, somebody is still reading, watching and listening. Those audiences are already assembled, already trusted, and mostly hungry for material. The fastest route to the first few hundred of the right people is to be useful inside somebody else’s room rather than to build your own from an empty list.

The forms that work are unglamorous: a teardown co-written with a practitioner your buyer follows, a podcast in the buyer’s function rather than in your industry, a newsletter placement where you supply the essay instead of the banner, a talk at the small conference whose audience does the job you are trying to change, a template given away with no gate in front of it.

What separates borrowing from spamming is whether the contribution stands on its own. If the piece is worth reading with your product removed, the host’s audience forgives the association and the host invites you back. If it is not, you get one appearance and a polite silence. The discipline that runs this is content and organic growth, not publicity, because the asset has to keep working long after the appearance ends.

Ask the host two questions before agreeing to anything. What does your audience complain about most, and what have you published on this subject already? The first tells you which symptom to write about. The second stops you offering a piece they ran last spring, which is the most common reason a pitch goes unanswered.

The asset you keep

Every borrowed appearance should end somewhere you control, and that somewhere should not be a homepage. A specific page, matching the specific thing you just talked about, carrying one request. The list you build from those pages is the only audience in this plan that cannot be taken away by an algorithm change or by a host who stops replying to email.

The sequence we would run

Order matters more than effort here. Doing these in the wrong sequence is how a well-funded launch produces a library nobody reads.

1. Write the problem sentence first. One sentence, in the buyer’s words, describing what goes wrong without your product. If three people in the company write it differently, nothing downstream will be consistent, and no volume of content repairs that.
2. Build the symptom library. A small set of pages answering the questions people already ask about the problem, written properly rather than quickly. These are the pages still earning attention long after the launch campaign is forgotten.
3. Take the adjacent queries. Comparison, alternative and do-it-manually pages. Lower volume, higher intent, and they produce sales conversations while the symptom library matures.
4. Borrow three audiences properly. Three real relationships beat thirty pitches. Give each one something that stands alone, and send its traffic to a page built for that audience rather than to the front door.
5. Instrument it before you need the numbers. Decide now how a lead that first heard of you on a podcast months ago will be recognised, because retrofitting that is miserable. Attribution you can defend is a design decision, not a reporting one.
6. Only then buy the category term. Once people are searching it, paid becomes a harvesting channel again and it is worth what it costs. Before that, it is tuition.

One operational note, because it is what quietly kills these programmes. Publishing on a rhythm for a year is an operations problem before it is a creative one, and the friction is almost never the writing. It is the formatting, the review and the deployment attached to every single piece. A structured workspace such as Acrosite removes that part: editors fill structured fields, it generates the files, commits them and triggers the configured deployment, so the cadence survives the week when everyone is busy.

The honest shape of the timeline

The curve is not a curve for a long time. It is a flat line with occasional spikes, each spike mapping precisely to a borrowed audience appearance, and each spike decaying afterwards. Then, at a point nobody can predict in advance, the flat line stops being flat, because the symptom pages have accumulated enough surface area to be found without a launch behind them.

The first honest signal is not traffic at all. It is language: prospects arriving in sales conversations already using your words for the problem, before anybody on your side has said them. That happens well before the analytics show anything, and it is the moment to keep going rather than to change the plan.

Our parent company has written about what ninety days can honestly show, and the shape it describes applies here with the volume turned down and the patience turned up.

The concession

This approach is slow, and slow is a real cost rather than a rhetorical one. If the runway is short, or the board expects pipeline this quarter, the correct move may be the opposite of everything above: describe yourself in the adjacent category’s language, take the demand that already exists there, and change the story once you have revenue. It is less satisfying and it works.

The price is that you will eventually have to re-teach a market that has already filed you under the wrong heading, and re-teaching is harder than teaching. That is a trade rather than a mistake. Make it deliberately, with a date attached to the point where you stop, instead of drifting into it because the first quarter was quiet.

Measuring a programme with no rankings to report

Something has to go in the monthly report, and rankings will not be it. The measures that mean something early are unglamorous and partly qualitative: the share of sales calls where the buyer used the problem language first, the number of inbound enquiries that arrive already understanding what you do, replies to the list rather than opens, and invitations to speak or contribute that nobody had to chase.

Cost per lead is a poor instrument at this stage. The earliest leads are cheap, unrepresentative and often the wrong people entirely, so a falling cost per lead can mean the message is landing or that it has started attracting bargain hunters. We have argued that cost per lead is a trap even in mature programmes. In a category that does not exist yet, it will actively push the team back towards whichever channel produced the cheapest wrong people.

Set the measurement plan before the first piece publishes. Retrofitting a definition of success onto a programme that has already run for two quarters produces an argument rather than a decision, and the argument is usually won by whoever brought the simplest chart.

A category is not a keyword you win. It is a sentence enough people can repeat when you are not in the room.

Where we would start on Monday

After that, three things in the first month. Find the three places your buyer already reads or attends, and write to one of them with something specific enough to be useful rather than a pitch. Write the one symptom page closest to the moment of purchase. Publish the definition page and then stop worrying about it.

The condition that reverses this advice is simple enough to check. If real search volume exists for the problem, even under a name you dislike, harvest it first and build the category later. Demand you can measure beats demand you have to manufacture, every time. If you are somewhere between the two, a product launch programme should be sequenced so that the harvesting pays for the manufacturing rather than the other way round.

And be suspicious of anyone who says this can be done quickly. The tactics are ordinary and widely known. The thing that is genuinely hard to buy is the willingness to publish into silence for long enough that the silence ends.