Grupo Casas Bahia
AI product search and recommendations for the marketplace
Grupo Casas Bahia lifted search conversion 58% after moving marketplace search and recommendations to Google's retail AI
Problem
Grupo Casas Bahia expanded its marketplace from thousands of items to millions. Its earlier search and recommendation tools gave poor result quality and low resilience, and customer conversion dropped.
Approach
The IT team ran a 45-day proof of concept with Google Cloud's product team and chose the Retail API for its three e-commerce brands. Retail Search, which learns from user trends to pick the sort order that maximizes revenue, went to Casas Bahia. Recommendations AI went to Ponto and Extra.com.br.
Results
Takeaway
The retailer judged the new tools on search relevance and model retraining as well as catalog update times. Search conversion, click-through and bounce rate moved after the switch.
Categories
- Size
- Large enterprise
- Industry
- Retail and e-commerce
- Business function
- Marketing and sales
- Innovation type
- Product and service innovation
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Google Cloud's customer story describes how Grupo Casas Bahia replaced its marketplace search and recommendation tools with Google's Retail Search and Recommendations AI. It reports a 58% rise in search-driven conversion, a 28% higher click-through rate, a 4% lower bounce rate and 28% more income per app user. The page carries no date; the Internet Archive first saved it on August 30, 2025.
Full details are in the original sources.