Product reality on one side. Buyer language on the other.
Reach Dog built an open-market commerce map that connects what products are to the ways buyers describe what they need. The map links product listings, semantic product neighborhoods, search language, natural-language buyer questions, attributes, five buyer dimensions, and market signals.
The map, layer by layer.
More than 360 million product listings
Product titles and descriptions provide the raw product reality the rest of the map resolves against.
29.6 million semantic product neighborhoods
Listings are grouped into comparable product neighborhoods by meaning rather than relying only on each merchant's category label.
70.5 million natural-language buyer questions
Full-sentence descriptions of needs, situations, outcomes, and constraints create a buyer-language layer that a category list does not contain.
These phrasings model buyer cognition. They are not presented as harvested query logs.
101.3 million analyzed keywords
Compressed search vocabulary adds demand and commercial context where those signals are available.
What. Where. When. Who. Why.
Product neighborhoods carry the dimensions buyers use to describe a need: what the product does for them, where it is used, when it matters, who it is for, and the benefit that makes the product useful.
Market context around the product
Pricing, competition, demand, and related market signals turn a semantic match into a commercial decision.
The connections are the asset.
A listing corpus alone is product data. A keyword corpus alone is search data. Reach Dog's differentiation is the map between product reality and buyer language, plus the ability to resolve a merchant's own catalog against that map.
A category tree and a commerce map solve different problems.
Conventional taxonomy
Organizes products into a standardized hierarchy so systems can classify what an item is.
Reach Dog commerce map
Connects products to multiple semantic neighborhoods and the buyer language around them, including uses, occasions, audiences, timing, benefits, and natural-language questions.
The difference shows up in live retrieval.
Reach Dog's search-mismatch study captured queries where buyers described situations and performance needs rather than naming a category. The study found repeated cases where editorial or AI could name the need while the commerce layer could not assemble the corresponding product set.
The captured Shopping filters included Type, Department, Features, Price, Age Group, Product Rating, Number of Speeds, and Tire Width. The buyer said "climbing hills." Zero observed product titles contained "climbing."
Work with the map at the level you need.
Research / data licensing
Contact Reach Dog for scoped data access and research use.
Contact Reach Dog →