Business Impact

  • ~47%

    Increment in Mortgage Purchase Model

  • ~25%

    Increment in Mortgage Refinance & Churn Model

Customer Key Facts

  • Location : North America
  • Industry : Financial Services

Problem Context

The client is an American bank holding company headquartered in Buffalo, New York. They wanted to build a mortgage profiling model that could identify customers who were likely to opt-in for a mortgage service. The bank had existing models that helped predict if a customer was likely to opt-in for mortgage services. However, those were using data from disparate sources to define customer propensity.



  • Stitching data across clickstream and internal data sources
  • High imbalance in data
  • Identifying relevant features for fair lending

Technologies Used

Google Cloud Platform
Cloud Functions
Cloud Storage

Built a New Machine Learning Model on Connected Data to Improve Overall Customer Targeting for Mortgage Services


After the bank moved all its data into GCS, Quantiphi migrated the data from GCS to BigQuery and stitched the data. Quantiphi leveraged feature engineering capabilities to prepare the data for modeling and built a new machine learning model on connected data that improves overall customer targeting for mortgage services. Our experts prepared three different models to predict mortgage purchase, refinance, and mortgage churn propensity. They mapped propensity scores to different user identifiers to enable multi-channel activation.


  • Achieved a higher conversation by targeting customers with a high probability of mortgage purchase or refinance
  • Ability to target customers across multiple digital & offline channels

Looking for similar project?

Let's Talk

Get your digital transformation started

Let's Talk