How to Improve Loan Origination Revenues with Champion / Challenger Testing

Credit scoring models have already become an essential part of loan origination systems. A growing number of lenders turn to Big Data technologies to set up a robust lending mechanism. John Thiel, Head of the Private Banking and Investment Group at Merril Lynch, says in his Forbes interview: “Big data is the ability to spot trends and opportunities that otherwise wouldn’t be available without the amount of information you can aggregate to do that…Being able to leverage the data enables us to help our clients think through implications and ways to mitigate and/or fund risks.”

Fierce competition and constantly changing economic environment challenge lenders to automate decisions in a way that learns to predict best outcome. Champion/challenger testing of credit policy enables the creation of prudent and profitable lending strategy.

How to evaluate changes in the credit policies
In order to develop the perfect loan origination strategy, risk managers need to choose the best way to pre-screen and qualify borrowers. They need to identify the optimal loan application processing flow, the appropriate way to select pricing parameters and more. Calculations, intuition and using past experience will not replace the value of modeling the effects of a given credit policy, or running a live experiment to compare performance of two credit policies.

With A/B testing, or Champion/Challenger testing approach, you can evaluate the effects of any change in the lending policy. You can proceed in the following way:

  1. Select the way of testing loan application processing strategies using a champion/challenger system.
  2. There are two possible options. First, you can model application processing in a virtual environment using data from your portfolio, or another statistical data. In order to test lending policy in a live environment, you should set your LOS to function in a champion/challenger mode and automatically process a certain proportion of applications using the champion and challenger lending policies.

  3. Design improvements in lending policy.
  4. Specify the changes in credit policy that need to be tested. For instance, set a different interest rate, adjust decision logic, improve scorecard accuracy, test a lower score cut-off parameter. Or you may want to follow advice that a known credit risk professional Brendan Le Grange gives in his article Credit Policy: How Much is Too Much: “broader policy rules makes people accountable and so they needn’t increase risk in any many cases they actually decrease it.”

  5. Identify the optimal credit policy and implement changes.
  6. Determine key performance indicators (KPI) according to your goals and business challenges. You may prefer monitoring operational indicators such as LGD ratio, delinquency rate, and average total time required to process a loan application. Alternatively, you can watch business performance indicators like proportion between approved and turned down loan applications.

Improving loan origination policy: best practices and recommendations
It is recommended to start testing the new credit policy by processing around 10-15% of borrower applications using the challenger policy. This will allow you to detect any possible flaws on time.

Keep in mind that changes in the lending strategies are best evaluated when testing a single improvement at a time. This will show you the exact effects of any particular enhancement.

About the author:
I am responsinble for marketing activities at Scorto. Here at Scorto we provide decision automation solutions for the financial services industry, telecommunications, and insurance companies. I am also maintaining our blog on risk management and decision automation.
My website is at:


Related Articles:

Leave a Reply