How Millennium bcp Uses BigQuery ML to Boost Loan Conversion Rates

Millennium bcp, Portugal's largest private bank, used Google Cloud's BigQuery machine learning tools to build predictive audience models for personal loan campaigns. By segmenting existing customers by propensity to borrow, the bank dramatically improved both owned and paid media performance. The result was a 2.6x higher conversion rate and a 36% drop in cost per acquisition.

Outcomes

2.6x higherConversion rate lift — owned media (BigQuery audiences vs. other first-party audiences)
2.4x greaterConversion volume lift — owned media (BigQuery audiences vs. other first-party audiences)
2x (doubled)Conversion volume lift — paid media
1.9x higherConversion rate lift — paid media
36% lowerCost per acquisition (CPA) reduction

Tools & Technologies

1F
Firebase
Mobile app engagement platform by Google Firebase for in-app messaging, analytics, and user retention.
2D&
Display & Video 360
Programmatic advertising platform by Google for managing and optimizing display and video ad campaigns.
3GC
Google Cloud
Comprehensive cloud platform by Google offering compute, storage, AI, and data services at scale.
4GB
Google BigQuery
Serverless enterprise data warehouse for analytics
5GA
Google Analytics 4
Web and app analytics platform by Google for tracking user behavior and marketing performance.

AI Categories

Challenge

Millennium bcp's reliance on in-branch processes for personal loan applications limited scalability and failed to meet modern customer expectations. The bank needed a way to digitally target only eligible existing customers with personalized loan offers at scale.

Solution

The bank's in-house digital marketing team built BigQuery ML-powered predictive models to segment existing customers by loan propensity, then activated those audiences across owned and paid channels via Google Analytics 4, Firebase, and Display & Video 360.

Full Story

Millennium bcp, Portugal's largest private bank by business volume, had long relied on physical branch visits for personal loan applications — a model that constrained scalability and failed to meet the expectations of increasingly digital-first customers. Clients were required to visit in person during limited hours and endure slow, manual processes. By 2023, the bank set an ambitious goal: meaningfully grow digital sales of personal loans while ensuring outreach was limited to existing customers, the only segment eligible to complete the digital loan journey.

Access 449+ AI use cases, 414+ tools, and adoption signal rankings.

Source

GOOGLE
March 2026
Original case study

Similar Cases

1K
How Klarna’s AI Assistant Resolves 80% of Queries in Under 2 Minutes
Klarna
80%Reduction in average customer query resolution time
2S
How Stripe Deploys Claude Code to 1,370 Engineers with Zero-Configuration Rollout
Stripe
1,370Engineers Deployed
3A
How Airtree Uses Claude Cowork to Automate VC Research & Reporting
Airtree
Reduced from 2 days to minutesMarket & competitor research time
4W
How WEX Achieved 30% Developer Productivity Gains with GitHub Copilot
WEX
~30%Developer productivity increase with GitHub Copilot
5S
How Super-Pharm Uses Vertex AI to Improve Inventory Accuracy from 50% to 90%
Super-Pharm
50% to 90%Inventory Accuracy
6F
How Fireblocks Uses Snowflake AI Agents to Handle 40-50% of Data Queries
Fireblocks
40–50%Share of data queries handled by AI agent
7NB
How NBIM Uses Claude Enterprise to Save 20% Time on Investment Analysis
Norges Bank Investment Management
20%Weekly time savings per employee
8B
How Block Gives 4,000 Employees AI-Powered Data Access via Claude and Databricks
Block
75% saving 8-10+ hoursEngineers saving time weekly
9S
How Shopify Scales AI-Powered Sidekick to Millions of Merchants with Claude on Vertex AI
Shopify
< 24 hoursModel Upgrade Deployment
10S
How Satispay Generates 75% of Its Code with Claude
Satispay
75%+Share of monthly committed code generated with Claude
See all use cases →