The problem
Search was already a strong channel—converting at roughly twice the site average. Growth stalled anyway. Shoppers typed the way they think about outfits and occasions: “airport outfit,” “quiet luxury workwear,” and other intent-rich phrases.
The catalog only spoke category, color, and basic product attributes. When the index lacked the vocabulary shoppers used, even a capable engine returned thin or mismatched results. People who knew what they wanted still could not express it in catalog language.
The approach
The retailer kept the same search engine and expanded the signal set feeding it. Perspiq generated contextual catalog fields and integrated them before indexing so relevance could match shopper intent—not just taxonomy.
- Contextual field generation — occasion, aesthetic, and use-case language attached to products
- Pre-index integration — new signals entered the pipeline before search indexing
- Signal expansion — catalog vocabulary grew to cover how shoppers actually search
The outcome
Intent-driven queries started engaging. Search-driven revenue rose 8% quarter over quarter, and shoppers needed fewer refinements to land on relevant results. The ceiling was not the engine—it was missing language in the catalog.
Search stopped failing shoppers who knew what they wanted but didn't know what to type.