Bank loan portfolios; Improving small business credit scoring effectiveness to grow performing loan portfolios

Project Details

Title: Incorporating unconventional data attributes to improve credit risk scoring
Customer: Major European Bank
Industry: Banking
Period of project: 2 Years
Budget: $1,000,000+
Est. value/ROI: 10 X - The scoring implemented as a pilot in several regions of the country increased the loan portfolio by 3-4% - Subsequent location rollout multiplied ROI
How many persons in project: 4
Used Technologies: Python, MATLAB
Hash-Tags/Keywords: Credit risk, Credit scoring
Project Category:
Risk insights

Over the past decade, we have become a data-driven, cloud-centric world.

For financial services companies, that transformation has been a double-edged sword. The banking industry is undergoing a radical shift, one driven by new competition from FinTechs, changing business models, mounting regulation, compliance pressures, and disruptive technologies. The emergence of FinTech/non-bank startups is changing the competitive landscape in financial services, forcing traditional institutions to rethink the way they do business.

Banks’ regardless of size and mission are sitting on a treasure trove of data. Visibility and electronic access to large amounts of data provide an unlimited opportunity for efficiency, risk mitigation, and customer service improvements leading to dramatic business models and industry transformation.

Our clients’ aspiration was to improve upon the traditional credit assessment and leverage data science to predictably drive more effective business loan performance than traditional methods alone.

Companies need a new approach to strategy for managing in a world of ecosystem disruption.

Automating repetitive tasks frees up a workforce to add value. Automating data analytics changes the field of play, the makeup of team members and overall dynamics. Automatically identifying relationships and patterns in data reduces the reliance on human expertise and judgment while keeping intentionality in the hands of a human actor. This is the essence of augmented intelligence using data science to make sense out of large sums of data in the context of a role and challenge enabling us to do more.

Challenge

Lower small business loan default rates, predictably grow loan portfolio and improve performance.

Solution

Leveraging unconventional transactional data and machine learning to assess credit reliability of small business applicants complementing and enhancing the Bank’s traditional credit decision-making.

Outcome

Initial impact >10 X ROI – The scoring implemented as a pilot in several regions of the country led to a 3-4% increase in the loan portfolio. The ROI multiplied as rollout and deployment commenced in each subsequent location.

Ready to Progress

If you’re curious to explore how data science can help you

  • solve complex challenges
  • drive efficiencies through automation, and
  • increase productivity by empowering your workforce to act with contextual insights

browse our services, and reach out to start a conversation

Solutions, resources, and bundled packages to help companies advance no matter where they are on their data science journeys regardless of industry or domain.

Together let’s take the steps to discover your first AI use case.

Take a look at how we helped a bank automate a data tape to reduce error-prone manual effort and deliver a consolidated risk picture improving investor relationship management.