The observation
Run Leiden clustering with Surprise optimisation on a real consumer category, and the segments emerge. Not 3 (the survey-research default). Not 5 (the agency consulting default). Almost always 8 to 15 for a focused product category, more for a broad one.
The Bose Germany headphone market resolved into eleven shapes:
- True wireless premium
- Over-ear noise cancelling
- Sport/wireless earbuds
- On-ear lifestyle
- Audiophile wired
- Studio/professional
- Kids headphones
- Gaming over-ear
- Bone conduction
- Sleep/ambient
- Budget-tier wireless
A 5-segment customer model would have collapsed seven of these into "other". The strategy that flows from "11 segments, here are the demand pockets" is fundamentally different from the strategy that flows from "3 segments, here is your positioning matrix".
Why legacy methodology finds 3-5
Three reasons:
01 — Survey statistics. Attitudinal surveys with 800-1,500 respondents can resolve maybe 5 segments before sub-segment sample sizes get too small for significance. The 5-segment output is a methodology artefact, not a market reality.
02 — Consulting communication. A 5-segment slide fits a boardroom narrative. 11 segments is harder to present and harder to action with a single positioning recommendation.
03 — No continuous data. Without continuous search-and-purchase data, you can't validate the 11-segment view. So nobody builds it.
Why search-data segmentation finds more
Search behaviour is continuous, at scale, and revealed preference. Every keyword is a consumer voting for a specific intent. The clustering algorithm only finds the segments that consumers actually generate by their searches.
If consumers don't behave as a segment, it doesn't show up. If they do, it does — regardless of whether the brand or the research firm previously thought it existed.
Categories and their shapes
| Category | Country | Shapes found |
|---|---|---|
| Headphones | DE | 11 |
| Mirrorless cameras | UK | 9 |
| Printers (consumer) | EU | 7 |
| Pro photo printers | EU | 3 |
| Baby/parenting (Graco scope) | UK | 14 |
| Luxury travel destinations | UK | 8 |
| Gift market (cross-vertical) | UK | 7 |
| Mortgages | UK | 6 |
| Beauty actives (retinol-led) | US | 12 |
Pattern: narrower categories produce fewer shapes (pro printers: 3). Broader categories produce more (parenting: 14). The total is always more than legacy methodology would propose.
What this changes about category strategy
Three actions follow from a richer segmentation:
- Identify your owned segments — where you have natural share advantage
- Identify your defendable adjacencies — segments where modest investment grows share
- Identify the segments to ignore — where competitors have structural advantages you can't overcome
Single-segment strategy is impossible without first knowing the segments exist. Eleven shapes is the precondition.