How to spot an emerging category in search data

Claire Taylor in Search Engine Land, 2026-08-06. Five signals that a market category is forming, read off keyword data before the competition arrives — with pulled figures from SE Ranking’s US and UK databases (July 2026) rather than assertions, which is why this is T2 in a spoke whose floor is mostly trade press.

Its subject is category-emergence, and it is the first source here that treats demand discovery as a method rather than a preliminary to keyword targeting.

The five signals

1. The authentic phrasing is invisible. “In an emerging category, the most authentic buyer language is invisible in keyword tools” — real buyer questions have not accumulated measurable volume yet, so the tools are structurally late.
2. Formal vocabulary moves first. Regulations, standards and job titles register before natural language does. Her numbers, 12 months: “AI governance framework” 40 → 3,600 (90×), “AI regulation” 120 → 3,600 (30×), “ISO 42001” 610 → 3,600 (6×), and “Colorado AI Act” from nothing in January to 760 by month 12.
3. The vocabulary is still contested. Governance, compliance, audit and risk describe overlapping things at similar volumes with no winner — “the vocabulary is still up for grabs.” A settled label means you are late.
4. Difficulty lags demand. The commercial signal: volume without competition. “AI governance consultant” 170 searches at difficulty 22; “data privacy consultant” 260 at difficulty 7; “AI policy template” 320 at difficulty 21.
5. The SERP is mismatched. Early results mix IBM, Accenture and the Big Four with two-person consultancies ranking above them — “relevance still beats authority,” which is exactly what stops being true once a category matures.

The discipline that makes it a method

Three rules separate this from trend-chasing, and they are the reusable part:

• Fixed keyword cohorts. Growth measured over a list you keep adding terms to is arithmetic, not demand. Most false positives come from here.
• A 12-month window minimum, to tell structural demand from a news spike.
• Commercial intent must appear. Informational queries without “consultant,” “template,” “certification” or “cost” behind them mean curiosity, not budget.

Her triangulation rule, which is the honest version of a soft method: “When the anecdotes and the data agree, trust the pattern.” The listening posts she pairs with the keyword data are TikTok and Instagram search, YouTube autocomplete and comments, Pinterest Trends, Reddit, podcasts and conference talks, plus first-party site search, sales notes and support tickets — on the argument that keyword tools lag social platforms by months.

The window, measured

The strongest evidence in the piece is geographic. The AI-governance cluster passes 25,000 monthly US searches and 12,000–13,000 in the UK, for a category that barely existed two years earlier — and US difficulty is already 68–72 where the UK equivalents are 25–32. The same category, two markets, the window visibly closing in one and open in the other. “The window doesn’t stay open for long.”

The consumer counter-example is Rise at Seven spotting “airport outfits” on TikTok before it had search volume, with PrettyLittleThing taking the category and ending at #1 on ~21,000 monthly searches and roughly 7,000 organic visits across nearly 400 keywords.

Where it sits, and what it is not

This is the demand-side counterpart to the spoke’s content-gap-analysis thread: that one asks what competitors cover and you don’t, this one asks whether the category itself is forming. It also gives the seo-operating-model-shift argument a concrete non-commoditizable task — noticing a market before the tools can price it is not work an AI content pipeline does.

What it does not do is answer this spoke’s standing measurement problem. Everything here is measured in searches and rankings; the open question about GEO is how to measure share of AI recommendation, and keyword difficulty says nothing about that. If the ai-search-shift thesis holds, the instrument this article sharpens is one aimed at a shrinking surface — worth stating plainly, because the article does not.

Cross-spoke, worth recording: the primary case study is AI governance, which this hub covers as its own spoke. Those figures are demand-side evidence that the field ai-governance-wiki documents (EU AI Act, ISO/IEC 42001, national strategies) is in early-adopter growth rather than mature. And the method itself is an empirical instrument for the diffusion/adoption theory research-wiki holds — a way to observe an adoption curve while it is happening, where that spoke has Rogers, Moore, the hype cycle and Bass as models of it.

Related

category-emergence · content-gap-analysis · seo-operating-model-shift · seo-tools · ai-search-shift · generative-engine-optimization · claire-taylor · search-engine-land · synthesis