The Knowledge Graph is the foundation of how Elevate understands products and search. It maps product types, attributes, and the relationships between concepts so that search results are based on meaning, not just matching words.
How it works
A structured map of products and concepts
At its core, the Knowledge Graph is a structured map of product types, their attributes, and how concepts relate to one another. For each product type, for example a shirt— a shirt, it defines which attributes matter: brand, shape, pattern, material, and so on. These aren't just labels. The Knowledge Graph understands that "slim" refers to a fit, that "cobalt" is a shade of blue, and that a bow tie and a tie are distinct products that belong to a shared contextual category.
The structure is hierarchical, meaning it captures both broad and specific ideas. "Clothing" is a valid concept, but "tank top" is a much more precise one. When a customer searches at any level of specificity, Elevate can connect their query to the right products, regardless of whether they search for something general or something very particular.
Multilingual by design
The Knowledge Graph also handles language. If your product catalog is in Swedish or German, customers can still search in English and get relevant results. This works because the Knowledge Graph contains multilingual dictionaries that map equivalent concepts across languages. The translation happens inside Elevate without having to depend on your products being described in the right language.
Continuously maintained and expanded
Two dedicated teams continuously maintain and improve the Knowledge Graph. One focuses on expanding and refining the product models; the other works on improving how individual customers' products are classified within them. It's a living system that evolves with new product categories, new terminology, and the needs of the retailers using it.
When applied to your catalog, the Knowledge Graph operates selectively. If you sell fashion, Elevate focuses on the parts of the graph relevant to fashion. If you're in tools and home improvement, it focuses there. The full graph is vast, spanning many industries and segments, but what matters in practice is the slice that applies to your products.
How to get the most from the Knowledge Graph
Keep your product data clean and consistent
The Knowledge Graph works best when your product data is accurate. The more consistently your catalog describes your products, with correct brands, clear titles, and well-structured attributes, the more effectively Elevate can map them to the right concepts in the graph.
Request expansions for specialist needs
For most retailers, the standard Knowledge Graph covers everything you need. However, some businesses have very specific requirements that the general model doesn't handle by default. If you sell licensed merchandise for particular sports teams, products tied to specific franchises or character brands, or operate in a highly specialized niche, the standard definitions may not go deep enough.
In those cases, reach out to your Customer Success Manager. The teams that maintain the Knowledge Graph can evaluate whether a custom expansion is appropriate for your catalog. For example, if a retailer sells merchandise for a specific sports league, Elevate can be extended to understand individual teams and return only the right ones and not just anything in the same general category.
Article last updated