There can be many reasons why banks and lending corporations fail. A study proposed by a research team led by Noura Metawa at the International Computer Engineering Conference (ICENCO) says the leading reasons are either due to an “unexpected decline in the value of the collateral used by the banks to secure the loans; an exogenous change in monetary conditions, (e.g. unexpected rise in reserve requirements or applying a new regulatory constraint that limits the lending process); the direct credit controls imposed by the government on the banking system; or an the increased perception of risk regarding the solvency of other banks.”
Which reason is most likely to hurt your bank or other lending institution?
During a financial crisis, it can be very challenging for banks and other lenders to know whether and how they can maximize their profits while distributing the limited credit available.
Thankfully, with new technologies and approaches, lending optimization can now harness data to provide valuable and powerful insights. These optimization solutions can help in each and every aspect of the lending experience – from testing to processing applications and beyond.
How can your bank use new technology to earn more revenue?
With perspectives from various banking experts, we list 9 possible optimization problems that you may run into and how to resolve them:
Hearing the Credit Crunch
The same study in the ICENCO says that one leading optimization problem is the inability of banks to manage loan portfolios efficiently, resulting in a credit crunch. A credit crunch happens when lending institutions become involved in careless and negligent lending, resulting in losses and debt when the loans turn sour.
Q: What is a credit crunch?
A: A credit crunch happens when lending institutions become involved in careless and negligent lending, resulting in losses and debt when the loans turn sour.
In order to circumvent these issues, they proposed a Genetic Algorithm (GA) as a self-organizing method to help make bank lending decisions and also solve other optimization problems such as credit assessment and portfolio optimization.
Genetic algorithms are a type of algorithm inspired by the natural selection process as proposed by Charles Darwin. GA mimics biological evolution and naturally adapts and evolves based on certain parameters and information that it receives. GA is used as a method for solving optimization problems.
Their proposed GA based model helps in optimizing bank objectives when constructing the loan portfolio. This in turn will maximize the bank’s profit and minimize the probability of defaults. Various data related to loan characteristics and creditor ratings are embedded into the GA DNA (and in turn, the genome). The GA then suggests the most appropriate design with maximum flexibility.
Getting the right people
According to Becky Franks, Digital Content and Optimization Manager of The Cooperative Bank, one of the greatest optimization problems in banking is having the right people to test different approaches and products. She says, “Getting the right teams to engage and drive the value of testing. People need to get into the mindset of testing and how it can help them reach their targets.”
For Becky, the best way to solve this is through prioritizing the testing backlogs with a clear hypothesis and benefits to the business and the customer. “If you have a prioritized backlog with data and insight to back up your test plans this really helps,” Franks says.
Long bureaucratic delays
Another optimization problem that Franks pointed out is working in a regulated environment, where all changes have to go through long approval processes. This causes a significant delay in overcoming challenges or meeting goals.
The only way to resolve this is to work with the system instead of trying to fight it. Through resilience, the frustratingly slow process of sign offs can be rewarding. But Franks also says to not be afraid of pushing back. “Just because you have been asked to do something doesn’t mean it’s the right thing to do, validate the data and if it doesn’t justify a test don’t do it.”
Low website traffic
When traffic to a bank’s website is low, achieving statistical significance can prove to be difficult. That is why it is important to always check your traffic volumes.
Heat maps can be a valuable tool when it comes to assessing and understanding where users are spending most of their time whenever they visit your website. These visual graphic representations of data can help you determine the “hot” spots throughout your site to let you know when, where, and why visitors discontinue and fall off in the conversion process. Heat mapping typically includes scroll depth tracking tools and tells you exactly where on your web pages a user stops scrolling.
The HiPPO effect
Q: What is a HiPPO?
A: The Highest Paid Person in the Organization/Room
David S. Bacon, Vice President of Digital Testing and Optimization of SunTrust Bank says that leaders, who primarily base their decisions on their personal experiences, may find conversion rate optimization challenging, especially with the staggering amount of data that most companies have at their disposal.
For David, the best leaders are comfortable with recommendations based on metrics, even if they don’t align with their own intuition. This means that a leader should learn how to analyze the data or hire someone who can.
Testing your ideas
Another challenge is turning testing ideas into actual valid tests. Ideas such as changing the color of the buttons to get more sales requires testing with careful methodology. Another challenge is that the tests may not result in the strategy that we envision.
According to David, always start with the data before beginning your tests. Use a framework like ICE to prioritize your a/b testing ideas to see what will likely have the most impact. Only after you get a clear direction to pursue should you design your tests. Always remember that data doesn’t support every testing idea that you conceive.
Confusing customer decisions and motivations
It is a common challenge among banks to deal with clients with very different motivations. Not all who visit the websites are going to behave the same way as the others. All customers have different intentions and exhibit different website viewing patterns.
Robert Martin, Vice President of Digital Channel Web Optimization at SunTrust Bank says that in order to not be overwhelmed by these varying amounts of data, one needs to dig deeper and only look at those that pertain to the area of focus. He says, “Further data discovery could show the underlying problem was actually people having trouble navigating to the checking page. This is somewhat of a simple example but highlights how different people’s motivations coming to a bank’s website could skew data.”
Choosing your favorite product to sell
Even the variety of products offered can pose a challenge when optimizing banking and lending websites. They can make traditional CRO (Conversion Rate Optimization) and A/B testing difficult. For example, one might see an increased conversion rate by 5% for a variant after running an A/B test on a website. While this may seem to be a winning experience, it might still not be profitable if the majority of the products sold were of less value — for example, savings accounts instead of credit cards.
Banks and other lenders need to understand how to balance the increasing volume and quality of accounts and the profitability of each. One way to help overcome this is understanding how many savings accounts would amount to 1 credit card.
Determining correctly priced loans
In these difficult times, it’s hard to achieve portfolio goals and significantly boost your bank’s profitability. But optimized loan pricing can greatly help.
Petr Kapoun, CRO of Home Credit Russia shares how his company decided to use decision optimization technology for loan pricing optimization.
Petr says that through decision optimization, guesswork can be replaced with science. Some of the key benefits of decision optimization include the use of analytics, rather than hunches, to simulate outcomes of changing strategies — as well as the ability to forecast business outcomes.