Keyword Research for a New Product Category (No Volume Yet)

When you’re doing keyword research for a new product category, standard tools won’t help you — Ahrefs, Semrush, and Google Keyword Planner all report volume based on past searches, and nobody has searched for your product yet. The fix isn’t to find hidden volume; there isn’t any. Instead, you research the problem your product solves, the analogous products people already search for, and the words your first 50 customers use to describe their pain, then build content around those proxies until real search demand catches up.

This is a different job than normal keyword research. You’re not mining an existing market for gaps — you’re guessing, with evidence, at a market that doesn’t have a name yet. Below is the process we use at PalV’s DM when a founder launches something genuinely new: a category-defining SaaS tool, a novel physical product, a service nobody’s packaged this way before.

What happens when your product doesn’t have a search category yet?

Search volume is a lagging indicator. It measures what people already type, which means it only exists after enough people already understand a problem well enough to name it. Airbnb didn’t get search volume for “vacation rental marketplace” until millions of people had already booked one. Slack launched in 2013 into a market where “team chat app” barely registered in Google Trends — the volume showed up two years later, after Slack had already spent two years teaching people what the category was.

If you wait for volume before you write content, you wait until your category is crowded and the keywords are expensive. The alternative is to write for the searches that already exist around the problem, not the product, so you’re ranking and building topical authority before the category keyword exists at all.

How do you find keywords for a product nobody searches for?

Four sources hold up in practice, in rough order of reliability:

1. Direct customer language. Pull the exact phrases from your first 20-50 sales calls, demo requests, and support tickets. Not your internal jargon — theirs. If a founder pitching a new expense-automation tool keeps hearing “I hate chasing people for receipts,” that phrase is a content topic, not just a quote.
2. Analogous product keywords. Find the closest existing category and mine its keyword set for terms that transfer. A new AI meeting-notes product can study “transcription software” and “meeting minutes template” — both have real, measurable volume and overlapping intent.
3. Problem-space keywords. These describe the pain, not the solution: “how to stop forgetting meeting action items” has search volume today even though “AI meeting memory assistant” has none.
4. Community and forum threads. Reddit, niche Slack communities, G2/Capterra reviews of adjacent tools, and Quora questions surface the actual words prospects reach for before a category name exists. Search “site:reddit.com [your problem]” and read the top 20 threads verbatim.

What is “jobs to be done” keyword research and how does it work?

The Jobs to Be Done framework, developed by Harvard Business School’s Clayton Christensen, reframes buying as “hiring” a product to do a job. Nobody buys a quarter-inch drill because they want a drill — they want a quarter-inch hole. Applied to keyword research, this means you stop searching for your product name and start searching for the job.

Take a genuinely new category: async video review software for design teams. Nobody searches “async video review tool” in meaningful numbers yet. But the job — “get feedback on a design without a live meeting” — already has searchers typing variations of “how to get design feedback without a meeting,” “async feedback process,” and “alternative to design review calls.” Those are your entry points. They’re lower volume than an established category term would eventually be, but they’re real, measurable, and winnable now.

Write down the job in plain language before you touch a keyword tool. One sentence: “helps [who] do [job] without [old painful way].” Every keyword you research afterward should map back to a piece of that sentence.

Which analogous products should you study for keyword proxies?

Pick two to four adjacent categories, not one. A single analogous category over-anchors your keyword set to a market that isn’t quite yours. For a new “AI-assisted contract redlining” product, you’d study:

• Contract management software — the closest functional category, high overlap in buyer intent.
• Legal document automation — adjacent workflow, shares the “reduce manual legal work” job.
• E-signature tools (DocuSign, PandaDoc) — same buyer, earlier stage of the same process, worth mining for “contract” modifiers.
• General AI writing assistants — different buyer but overlapping “AI does my drafting” search language, useful for top-of-funnel awareness content.

Run each analogous category through a real keyword tool and export volume, difficulty, and the “also rank for” list. You’re not targeting these keywords head-on — you’re stealing their modifiers (“free,” “template,” “checklist,” “vs,” “for startups”) and testing which ones make sense attached to your problem-space terms.

How do you use problem-space keywords instead of product keywords?

Split every keyword candidate into one of two buckets: product-space (names the thing) or problem-space (names the pain). Early on, your content plan should run 80/20 in favor of problem-space, because that’s where the volume already lives. Flip that ratio over 12-18 months as your product name starts accumulating branded and category searches of its own.

What does a category-creation keyword map actually look like?

Here’s a simplified version of a real mapping exercise, using the async-video-feedback example above:

*Illustrative ranges based on typical adjacent-category patterns, not a specific tool pull for this example — always confirm with a live check in Ahrefs, Semrush, or Google Keyword Planner before publishing.

Notice the pattern: volume climbs as you move away from your actual product and toward the established, adjacent language. That’s normal. You publish across all five rows, weighted toward the middle three, and let the product-space and job-to-be-done rows catch up as your category matures.

How long before real search volume shows up?

There’s no fixed timeline, but the pattern repeats across category creators: 18-36 months of near-zero branded and category volume, then a visible ramp once a critical mass of buyers has been through a sales cycle and started describing the product to peers using a shared name. Slack, Zoom (before “video call” became generic for it), and Notion all show this shape in Google Trends — flat for one to two years post-launch, then a steep climb once word-of-mouth caught up to the product.

Until then, this isn’t wasted content. Problem-space and analogous-category articles rank on their own merit, bring in the right visitors, and get you found by AI answer engines that summarize “how do I solve X” queries — which matters more every year as more of that top-of-funnel research happens inside ChatGPT and Google’s AI Overviews instead of ten blue links.

What tools help with category-creation keyword research?

This entire process sits inside our full keyword research process — category creation is just the edge case where step one (pulling volume from a keyword tool) doesn’t apply yet, so you lean harder on the qualitative steps. If you haven’t nailed down your starting keyword list method, start with how we build seed keyword lists before layering the category-creation techniques on top.

Two more pieces worth reading alongside this one: our breakdown of why zero-volume keywords still deserve a content slot, which covers the general case beyond new categories, and how keyword difficulty scores work — useful once your category term starts registering and you need to know if it’s worth chasing yet.

FAQ

Can you do keyword research if a keyword tool shows zero volume for everything?

Yes. Zero volume in Ahrefs or Semrush means the tool has no historical search data, not that nobody is searching. Shift to problem-space and analogous-category keywords, which do have volume, and treat your actual product term as a placeholder you’ll optimize for once real searches start.

Should a new product target its own category name as a keyword?

Include it on product and homepage pages for when volume arrives, but don’t build a content strategy around it. A category name with 0-10 monthly searches won’t carry a blog, a comparison page, or a landing page on its own for the first year or two.

How many analogous categories should I research?

Two to four. One anchors your keyword set too narrowly to a market that isn’t quite yours; more than four dilutes the research and slows you down without adding much new signal.

Is Jobs to Be Done the same as buyer personas?

No. Personas describe who the buyer is (title, company size, industry); Jobs to Be Done describes what they’re trying to accomplish, independent of who they are. For keyword research, the job matters more since two very different personas often type the same problem-space search.

When should I stop writing problem-space content and start writing product-space content?

Track branded and category-term impressions in Google Search Console monthly. Once your category term clears roughly 50-100 monthly impressions organically, start layering in product-space content alongside the problem-space articles rather than replacing them.

Does this approach work for a genuinely new physical product, not just software?

Yes, the same logic applies. A new kitchen gadget with no category name yet still has a job people already search for in relation to existing tools. Study the closest adjacent product category’s keywords the same way you would for software.

Written by Palash, founder of PalV’s DM. 5+ years in SEO, 1,000+ articles published. If your product doesn’t have a keyword yet, our one-time keyword research package maps the problem-space and analogous-category terms above for your specific market, delivered as a single project with no monthly retainer.