Customer story
How Kilte made their beautiful catalog searchable
Kilte's product descriptions are evocative and on-brand. "Cappuccino," "Ivory," "Midnight." The problem: shoppers don't search that way. Nobi bridges the gap, and beat Shopify search by 21.7% in the process.
About Kilte
Kilte is a fashion brand selling directly to consumers online. Their catalog leans into evocative, editorial naming — the kind that builds a brand identity and tells a story. Colors have names like Cappuccino, Ivory, Écru, and Midnight. It looks beautiful in a lookbook. It's a problem in a search bar.
The problem
Shoppers search for "brown boots," not "cappuccino boots." Keyword search has no way to know that Cappuccino is brown, or that Ivory is close enough to white, or that Midnight is black. It can only match what the shopper typed to what's in the product record. If those words don't match, the shopper gets nothing.
What Nobi does differently
AI-powered color normalization
Nobi builds a semantic layer on top of Kilte's catalog that maps brand color names to the standard colors shoppers actually use. "Cappuccino" becomes searchable as brown. "Ivory" becomes white. "Midnight" becomes black. When a shopper searches "brown boots," Nobi knows to surface the Cappuccino Boot even though the word "brown" never appears in the product description.
Cappuccino Boot not surfaced. Sale at risk.
The result
21.7% more conversions in a direct A/B test
Kilte ran a straightforward head-to-head: Nobi against Shopify's default search, measured on conversion rate. The result was a 21.7% improvement. They didn't need more convincing. They moved to Nobi across their entire site and haven't looked back.
See what Nobi can do for your brand
Set up takes minutes. Results show up in weeks.