Skip to content
Search Intelligence 2 min read

What Search Language Reveals About Category Structure

Merchant taxonomies describe how retailers shelve products. Search language describes how shoppers think. Where the two disagree, you'll find assortment, naming and SEO decisions waiting to be made.

Retail taxonomy is a institutional artifact: committees, legacy systems, supplier agreements. Shopper language is nobody's artifact — it's just what people type. Mapping one onto the other is one of the highest-yield exercises in category intelligence, because the mismatches are actionable by construction.

The method in brief

  1. Harvest the demand language: query clusters and modifiers around the category from aggregate public search signals.
  2. Express the merchant taxonomy in the same units — each branch as a term set.
  3. Align clusters to branches; flag demand that maps nowhere and branches that attract no demand.
  4. Validate flagged areas against catalog data — is the gap real inventory, or just naming?

Three mismatch patterns, three decisions

PatternWhat it usually meansDecision it feeds
Strong query cluster with no taxonomy branchDemand the catalog doesn't speak toAssortment expansion or naming audit
Deep taxonomy branch with flat demandMerchant structure shoppers don't usePage consolidation; taxonomy simplification
Cross-branch modifier ('for X' spanning two trees)Shoppers combine attributes your structure treats separatelyFaceted navigation and content strategy

The third pattern is the most consistently valuable. When shoppers habitually cross two branches the org chart keeps separate — "waterproof" across shoes and hiking gear, "refurbished" across laptops and tablets — the crossover is a product-design signal hiding in plain sight.

Demand mapped vs. unmapped by merchant taxonomy (share of query volume)
64Mapped22Partial14Unmapped
Source: Heroku taxonomy audit — share of monitored query volume. Unmapped demand above ~10% usually repays a naming audit.

Naming is positioning

The vocabulary shoppers use encodes what they value — "quiet" vs "low-noise", "starter" vs "beginner". Product pages that echo demand language convert better not because of SEO magic but because recognition precedes trust. The cheapest intelligence in this entire post: pull the top modifier lists for your category and read them next to your product titles. The gap is usually audible.

Your taxonomy is how the company thinks. Search language is how the market thinks. Products live in the difference.

— Elena Rossi

Run this exercise quarterly, alongside your demand-trend reviews, and the taxonomy stops being furniture and starts being an instrument.

Elena Rossi

Elena researches search behavior and category demand. She helps product and content teams understand how people actually describe what they want, and what that means for assortment and positioning.

Related reading

Continue here

3 comments

Derek Wong

Ran the four-step method on our category last weekend. Unmapped demand was 19%. The naming audit starts Monday.

Elena Rossi

19% is high enough that the naming audit will likely pay for itself quickly, Derek. Good hunting.

Ingrid Novak

“Your taxonomy is how the company thinks. Search language is how the market thinks.” Framed. Literally.

Join the discussion