Predictive Analytics for Consumer Lending: Introduction


The consumer lending business is centered on the notion of managing the risk of borrower default. Credit scoring systems and predictive models attempt to quantify uncertainty and provide guidance for the identification, measurement and monitoring of risk.

A vast majority of lending institutions (over 90%) use FICO scores when determining creditworthiness.1 The FICO score breaks down an individual’s financial profile into five categories: payment history (35%), outstanding credit (30%), length of credit history (15%), amount of new credit (10%), and credit types (10%).2 By considering a weighted index constructed from these categories, the FICO score is able to provide a picture of an individual’s creditworthiness.

What exactly does Predictive Analytics have to offer?

Predictive Analytics is a toolbox that includes mathematical techniques and processes that are applied to historical data to study correlations, identify trends and make predictions on possible outcomes by quantifying the uncertainty and the characteristics of the variation.

In the case of identifying creditworthiness, a data scientist is tasked with finding the right combination of factors that can most accurately explain an individual’s risk profile. These factors could include the individual’s financial profile, current economic measurements (GDP, GNP, etc.), geographic data, etc.


Data is everywhere!

Being in the midst of the “Big Data” age has it perks. The volume of data leaves an opportunity for organizations to become more educated and engaged about their business.

It’s never a bad time to build a better business.

Increase revenues and profits while reducing cost through improved risk management and increased operational capacity.

The current environment is perfect for growth.

As the country finds its way out of the financial crisis, purchasing power of individuals will rise. The battle for credit-worthy consumers will follow.


As predictive analytics continues its ascent to mainstream popularity, evidence has shown a gradual move toward credit scoring strategies developed using data mining and predictive analytics. The benefits associated with these predictive models include:

  • By taking into account information beyond the individual’s credit report, a more accurate estimation of a borrower’s default risk can be calculated.
  • Lending institutions are able to reduce operational cost and operational risk, as the loan application life cycle can be completed with fewer individuals performing fewer steps.
  • Predictive models make consumer lending decisions easier and improve the overall quality of the borrower portfolio, bringing in consistency and predictability in the loan-disbursement process.
  • Reduced bias. Automated credit-scoring models calculate the risk using the lending institution’s own history along with the borrower’s financial statistics resulting in an objective credit risk evaluation with reduced bias.

The new wave

Earnest Financial (, a San Francisco startup, uses predictive analytics for what they call merit-based lending. They use data, drawn primarily from LinkedIn, to provide an assessment of borrowers that is beyond what their financial profile reveals. By taking into account educational and professional background along with individual financial data, they are able to provide a more comprehensive picture of an individual’s risk profile. Neo (, makes use of the strength of a potential borrower’s social media connections to help lenders determine creditworthiness.building1

The emergence and popularity of these new players is a testament to the efficacy of predictive models as a viable business proposition. Their success can be attributed to the increasingly efficient processing and transmission of information today. As the volume of data continues to increase, it seems fair to conclude that predictive analytics has a viable place in the lending industry.

In the next article, Predictive analytics for consumer lending: From insights to decisions, we will discuss the use of predictive models in the context of the consumer lending business.

1 According to, “FICO Scores are the standard credit scores in the US, used in more than 90% of lending decisions.”

2 According to, these are the factors that affect the FICO score along with their relative weights.



Leave a reply

Your email address will not be published. Required fields are marked *