The science of
revealed reality.
How we turn what consumers actually do — not what they say they do — into structural certainty for the boardroom.
Revealed over stated.
Always.
The fastest way to get a category wrong is to ask consumers what they think. Stated preference is filtered through ego, fashion, and what people believe they should say. Revealed preference is what they actually do — what they search for, watch, buy, and complain about afterwards.
Urchin works only with the second. We don’t run surveys. We don’t field panels. We map the deep floor of the category — the heavy preferences that move slowly and matter most — by observing behaviour at scale.
We bypass stated opinion to map the deep floor of the market — the structural truths underneath the surface chatter.
We isolate the slow-moving, heavy-weight preferences that define long-term category shifts — not fleeting trends.
Urchin observes what consumers actually do — across search, content, conversations and commerce.
An omnidirectional sensor. Millions of data points across four behaviourally distinct signal types — each capturing a different layer of what the consumer reveals.
What consumers actively look for and aspire to.
Verified search demand across more than two billion queries. Long-tail problem statements, head-term aspirations, brand consideration, transactional intent. The most honest signal there is — search is what people type when no one is watching.
What consumers watch, learn, and adopt.
Native search across video and creator platforms, read on rolling four-week velocity windows. High-engagement themes, creator influence, brand visibility in cultural discovery. What consumers consume here predicts what they search for next.
What consumers openly discuss and question.
Forum and community surfacing through advanced search operators, plus open-text mining of review platforms. Unfiltered consumer language, pain points, brand perceptions — the questions people ask peers because they won’t ask brands.
What consumers experience after the purchase.
Review and rating signals at scale, alongside price-comparison search behaviour. Satisfaction drivers, deal-breakers, repeat-purchase markers. The signal that closes the loop — what people say after they’ve spent.
One signal can lie. Four can’t.
Five metrics. One picture of the category.
Raw data is a liability. We translate the four signals into five calibrated metrics that together describe a category’s commercial physics. Each is benchmark-calibrated across categories — so “high readiness” means high against a defined universe, not vibes.
Demand Index
How big is this category, in absolute search volume?
Not raw search volume, not GMV, not ad spend — the calibrated, intent-weighted volume of what consumers genuinely want, benchmarked across categories so it stays comparable.
Engagement Readiness
Are consumers in research mode or buying mode?
Of the people in this category, what proportion are trying to fix an urgent problem right now versus just browsing? A high-readiness category converts faster. A low-readiness category needs education.
Demand Spread
Concentrated in head-terms or distributed across long-tail?
Is this category driven by one hero need, or by many distinct consumer problems? Concentrated spread means a single message can win it. Wide spread means the category is fragmented and needs portfolio thinking.
Demand Over Time
Is the category growing, plateauing, or contracting?
Combines current size with long-term trajectory. Surfaces three distinct strategic shapes: large and growing (invest), mid-size and accelerating (strategic bet), large and stagnating (defend).
Buying Intent Index
How decisive is the audience once they show up?
Readiness against Spread. How decisive is this category’s audience relative to how fragmented it is? A high index means even a small audience converts hard. A low index means the funnel needs work.
The picture that ends every Urchin report.
Five metrics collapse into one chart: Buying Intent Index on the horizontal, Demand Over Time on the vertical. Every category in your set lands somewhere on this grid. The quadrant tells you what to do.
Lead with these.
High scale, decisive audience, accelerating demand. Defend market share aggressively and price for premium positioning.
Hold the line.
Large category, lower decisiveness. The audience is in research mode — fix the funnel before competitors do.
Place small bets.
Small base, decisive audience. Niche but high-margin — ideal for premium sub-brands or specialist plays.
Walk away (mostly).
Structurally weak: small, fragmented, decelerating. Re-enter only if there is a specific structural reason to play.
Intelligence designed for the boardroom — a balanced, reinforced view where every signal validates the next.
Five-fold synthesis.
No signal stands alone.
Search tells you what people want. Content tells you what they’re being shown. Conversations tell you how they’re talking about it. Commerce tells you what they thought afterwards. Behavioural science tells you why. A senior analyst layer makes the call on what it all means.
Five sources. One architecture.
Geometrically sound. Impossible to tip over.
What separates intelligence from data.
One category, read twice. What people say when asked, against what they reveal when nobody is asking.
“I buy on value for money.”
“I switched because of a rash.”
Surface versus depth
What consumers say they want, against what they reach for in the dark. Real decisions are faster, more emotional, and routinely contradict the survey response.
What actually drives choice
In baby diapers, price-value ranks third among category drivers and per-piece arithmetic is 80.6% of all price-attention signal — yet under 0.2% of records name an actual price. 61% of category entries begin with an overnight leak; 28% with a rash.
Who actually decides
Reviews celebrate the involved father at length and unprompted. He is 1.8% of role-cued voice and measurably never changes the brand or the size. The mother is 92.6%, and she decides alone.
A survey tells you who says they are involved. Revealed data tells you who signs off.