+14.72%

Increase in Purchase Conversion Rate

+19.84%

Increase in Search Revenue per User

Industry:Jewelry & Accessories

Business Unit:E-Commerce

Use Case:Product Discovery & Conversion Optimization

Channel:Website

Key Objectives:Enhance Product Relevance, Streamline Shopper Experience

Products:

Fine JewelryPersonalized Pieces

How Mejuri Increased Purchase Conversion by 14.7% with Marqo

Company Overview

Mejuri is a Toronto-based fine jewelry brand that redefined "everyday luxury" by selling high-quality, modern designs at accessible prices through a direct-to-consumer model. Operating across ecommerce and a growing retail footprint, Mejuri manages a dynamic catalog of fine jewelry and personalized pieces.

Mejuri identified product discovery as a strategic growth lever within its ecommerce experience. As the brand expanded its assortment and introduced new collections, search became increasingly important in shaping how customers explore styles, evaluate materials, and confidently complete high-consideration purchases.

Jewelry discovery requires precision. Shoppers search using nuanced combinations of metal type, stone, size, occasion, and personalization attributes, often expressed through natural language or descriptive phrases. The objective was to establish a scalable, AI-driven foundation for search that continuously improves performance and unlocks incremental revenue.

Why Mejuri Evaluated Marqo vs. Algolia

Mejuri needed a solution that could move beyond keyword relevance and deliver consistent, catalog-aware results across a growing assortment. After evaluating Algolia alongside other alternatives, Mejuri selected Marqo for its ecommerce-built foundation, adaptive ranking capabilities, and ability to understand nuanced, style-driven queries without heavy manual configuration.

The decision was guided by Marqo's ability to continuously improve relevance from real shopper behavior — reducing reliance on manual merchandising rules while keeping business control firmly in the team's hands.

Challenges

As Mejuri's catalog expanded, the limitations of keyword-based search became increasingly apparent. Shoppers were searching in natural language — using terms like "minimal gold stackable ring" or "dainty diamond pendant for everyday wear" — and the existing system struggled to surface the most contextually relevant products.

Search Quality and Relevance — Descriptive and style-driven queries did not consistently surface the most contextually relevant results, reducing customer confidence at key purchase moments.

Ranking and Personalization — Limited session-level adaptation meant product rankings did not reflect individual shopper preferences or real-time behavioral signals.

Operational Overhead — Sustaining performance required ongoing manual merchandising and rule adjustments, creating operational overhead across the team.

Visual and Attribute-Aware Discovery — No scalable mechanism to support visual similarity or attribute-aware ranking without extensive manual tagging and configuration.

Jewelry discovery requires precision — shoppers use nuanced combinations of metal type, stone, size, and occasion, often expressed through natural language. The objective was search that understands how customers think, not just what they type.

Solution

Mejuri deployed Marqo across its ecommerce search to bring adaptive, catalog-aware relevance to its fine jewelry assortment. Marqo's semantic understanding improved results for style-driven queries involving metal type, stone, size, and occasion — without requiring manual synonym lists or taxonomy management.

Behavioral signals from real shopper interactions continuously refined rankings over time. The Mejuri team retained full merchandising control throughout, with visibility into how onsite signals influenced relevance across collections and new arrivals.

Enhanced Semantic Understanding: Improved interpretation of natural language and style-driven searches, reducing inconsistent results and null queries across Mejuri's fine jewelry catalog.
Adaptive Ranking: Product rankings continuously adjusted using real-time behavioral signals, preserving merchandising control and transparency for the Mejuri team.
Catalog-Aware Relevance: Stronger alignment between search results and key product attributes such as metal type, stone, size, and personalization options.
Reduced Operational Overhead: Reduced reliance on manual rule tuning and engineering resources as rankings improved through learning from anonymized onsite behavior over time.

Results

Mejuri evaluated Marqo using a controlled on-site A/B test with 50% of traffic allocated to Marqo. During the test period, Mejuri observed significant gains in conversion and monetization.

Overall Business Impact

+14.72%

Increase in Purchase Conversion Rate

+19.84%

Increase in Search Revenue per User

What This Means

The introduction of contextually aware, adaptive search improved discovery quality and directly translated into stronger purchasing intent. For a fine jewelry brand where shopper confidence is everything, surfacing the right piece at the right moment drives measurable conversion gains.

The right jewelry surfaces for the right intent — driving confident purchase decisions.
Rankings adapt automatically, reducing manual merchandising overhead over time.
Marqo now powers browse and recommendations across Mejuri's full discovery stack.

Results reflect a controlled A/B test conducted on Mejuri's production environment. Outcomes may vary depending on catalog size, traffic volume, and implementation configuration.

Moving Forward

What Comes Next for Mejuri

With positive results from search, Mejuri is expanding Marqo to power browse pages and recommendations across its site. Marqo is now a core component of Mejuri's discovery stack, enabling a scalable, intent-driven shopping experience that reflects how customers think, search, and buy fine jewelry online.