A Strategic Framework for Improving Loan Approval Rates
Vision and Strategy Document
Authored by: Bridgeforce, Inc.
The challenge that all marketers face in acquiring and retaining “the right” new customers is a daunting one. With the continued commoditization of loan products and the widespread acceptance of scoring and automated lending programs, loan marketers face some fierce competition. It is not enough to have the lowest rates, innovative products, or advanced marketing strategies these days. Loan marketers need to make sure that once they have a customer’s attention they don’t inadvertently loose an opportunity by making a credit decision based on inaccurate and/or limited information.
By applying the basic principles of the strategic framework outlined in this white paper, progressive financial service companies can gain a competitive advantage by identifying creditworthy customers who may be slipping through the cracks. Allowing lenders to proactively identify and resolve potential weaknesses within; prospect targeting criteria, manual credit review criteria, credit policies, and line assignment / loan amount policies and ensure that they are getting the maximum return on their marketing program investments (ROMI).
Progressive financial service companies that are prepared to invest the time and effort required to fully optimize their new customer acquisition programs, will gain the type of competitive advantages that were only available to a select group of lenders before the widespread acceptance of scoring and automated lending processes.
Identifying New Opportunities
Today’s lending industry is much like the lumber industry
during the 1980’s. Two decades ago,
lumber yards had large old-growth trees from which they could mill premium
boards. Scrap wood, bark and sawdust
were burned as waste. As old-growth
trees became scarce, the lumber industry had to find new ways to generate
revenue. They finally realized that they
were literally letting their future go up in smoke. From their waste, they devised new products.
For example, wood chips became oriented strand board, saw dust became a
component of many manufactured building materials and fuel pellets, bark became
mulch, and inferior boards became the raw material for engineered beams. The waste products of yesterday’s lumber
industry are now the foundation of today’s new construction materials and
In our view, the lumber industry is a metaphor for the lending industry of tomorrow. The “waste” of today’s lending industry is unattended credit needs. In order to survive and grow, lenders will need to look beyond just the “premium” applicants that everyone else is targeting.
Simply put, it’s time to put some of the art back into the science of lending!
The widespread move toward automated lending and scoring processes has allowed lenders to take advantage of lower loan processing costs, handle far greater volume, and improve decision consistency levels. However, for a variety of reasons, some of the applicant populations that may have been approved through manual credit reviews are now being underserved.
The power and success of scoring is not being challenged here. However, it is important to remember that scores are both developed and executed from data, and as such, they are only as good as the data used to develop them. A score that uses application data is an example of this. Development data is likely to have far fewer observations of high income applicants. Due to limited observations, it stands to reason that the score for applicants in the higher and lower ends of the distributions would not be as powerful. Therefore, people who earn $1,000,000 could be treated the same as people who earn $100,000.
Similarly, data entry errors such as an applicant’s interpretation of application data fields or credit bureau data may not be identified in many automated lending processes. A recent study completed by the Public Interest Research Group (PIRG) revealed the following insights into potential credit data quality issues:
Through Bridgeforce’s ongoing research and analysis, we have found that a high percentage of missed new customer acquisition opportunities can be found in the following areas:
Missed prospects due to non-optimized targeting processes
One often overlooked side of the new customer acquisition equation is the effect of prospect targeting on approval rates. Some creditworthy prospects may be overlooked by outdated prospect quality exclusions. One such exclusion is “Number of Inquiries”. Due to industry changes, these exclusions may no longer be as indicative of risk. Another common exclusion for self-employed or small business applicants is Time in Business. While it is a known fact that an estimated 80% of all new businesses fail in the first 1-2 years, lenders who exclude all businesses open less than 2-3 years are leaving too much future business potential on the table. Reviewing other criteria such as; business type, personal credit, financials/investments, and the time the owner has been in a specific field of business could provide greater insight into identifying the lower risk 20% in this group. Consequently, to find more potential approvals, exclusion criteria should be continually reviewed and revised as necessary.
Product mix should also be a consideration in the prospect targeting criteria. Some institutions exclude prospects, not because of risk, but because they are not profitable. Development of more appealing products can help make high scoring prospects profitable. Likewise, risk-based pricing can allow loan marketers to expand their prospect populations into new score ranges.
Declines due to non-optimized automated lending processes
As with all automated lending processes, careful analysis is required. Waterfall reports should be generated to determine where applications end up and why. In particular, declines should be reviewed and the business group should not only understand what rules are driving the declines, but they should also determine if the rules need to be modified. One frequent offender is data entry and/or application data field interpretation errors. As it relates to income hurdles, these errors can have a serious impact on the decision outcomes; debt to income ratios and line assignments (loan amounts) may be off by significant amounts. The business group should identify where data entry and application errors occur, and then work to minimize them. To identify applications that are at risk for these errors, business rules should be generated and the applications should be sent for a secondary review to verify application data that does not make sense.
Declines due to credit policies
credit policies have univariate (single characteristic) exclusions. Some are well founded in that they exclude
prospects and applicants that will most likely not be approved. However, when detailed analysis is not
can result in the elimination of creditworthy applicants. For example, common elimination criteria are
public records or collection accounts. Many lenders continue to eliminate
applicants with public records or collection accounts without first taking into
account their severity level or status.
The problem is, many creditworthy people do not know that they have
public records or collection accounts, in addition to the fact that these types
of accounts can be slow to update even when they have been resolved. Other common univariate exclusions include:
the number of inquiries, various regions, income, number of trades, and
more. To allow for the development of
more robust multivariate exclusions, we recommend slowly testing univariate
exclusion segments before eliminating or extensively modifying credit policy.
Declines due to non-optimized manual credit review processes
Manual credit review processes can add significant value to automated lending processes when credit analysts are used to clarify or obtain additional information from applicants. However, the real value in manual credit reviews is the feedback loop that is created once credit analysts obtain additional information and make a decision. Lenders who take the extra step to accurately capture the feedback provided to continually refine their scoring strategies, referral populations, and lender guidelines will ensure continual improvement in their automated lending processes.
Adverse Selection due to inappropriate line assignment or product offerings
Adverse selection is a situation where targeting strategy, application processing, credit policy, or customer service inadvertently biases the customer population toward less creditworthy customers. Adverse selection is important because it specifically discourages higher creditworthy prospects and responders in favor of lower creditworthy customers.
The following are examples:
Application Data Accuracy Strategies
Most organizations perform periodic reviews of applications in order to determine the quality of the input data. The following strategies will increase the quality of the application data prior to entering automated lending systems:
· Copy Tests – Test different credit application copy to determine which fields yield higher data accuracy rates.
· Automated Decision Rules – Ensure that automated rules are in place to capture potential application data errors and refer them for secondary review.
· Automated Data Entry – Utilize automated data entry processes and/or online applications, eliminate the potential for data entry errors by vendors or in-house processes.
Credit Bureau Data Accuracy Strategies
While the quality of data from the credit bureau agencies continues to increase, errors still exist that can impact the accuracy of credit decisions made. For example, “No/Thin Files” represents one of the largest populations of missed opportunities. Analysis should be performed to quantify the “Valid No File”, “Valid Thin File”, “Questionable No File”, and “Questionable Thin File” populations. Lenders should develop products appropriate for the valid populations, and route the questionable populations for additional development.
Manual Credit Review Strategies
Lenders can leverage manual credit reviews as a strategic opportunity to validate or confirm the information provided by applicants. For example, applicants that would be expected to have a robust file but don’t, may be able to provide corrected Social Security Numbers. In many automated lending processes, these applicant’s would receive, at best, a letter explaining why they were turned down. If the organization is lucky, a small percentage of such applicants may provide corrected information that could lead to an approval. Similar opportunities also exist for mixed files, mismatched files, or files with conflicting debt and income information.
Line Assignment / Loan Amount Review Strategies
Lenders can conduct manual line assignment / loan assignment reviews to challenge existing strategies. These should be limited in volume and targeted to specific populations where the automated process is more likely to assign non-competitive lines or where data errors are more likely to exist. Results from these reviews should be analyzed and used to further “tweak” the automated process.
As the high-scoring “premium” credit market becomes increasingly saturated, organizations want to develop offerings that meet the needs of the applicant as well as the issuer. This could mean offering a higher priced product or allowing another institution to underwrite the loan to satisfy a customer’s needs (especially if they have an existing relationship with your organization).
Marketing, Credit Policy, and Finance groups need to begin to look differently at previously declined score ranges. In many organizations, the first response will be “NO” when asked to consider ways to approve previously declined responders. They will want to have data to support assumptions about loss rates and profitability, but the data does not exist for new populations. Cautious champion / challenger testing is the way to go. These expanded populations and new products need to be carefully tested on small sample populations and the results tracked through the use of vintage delinquency reports before rolling out larger scale programs.
Credit analysts must have the ability to objectively consider all information and be able to identify what needs further clarification in order to assess the applicant’s true creditworthiness. Many leading lenders invest months in the training process to develop a credit analyst’s investigative phone skills. After their initial training, they provide further on-the-job developmental training to ensure that credit analysts continue to develop their lending skills in line with the organization’s lending guidelines.
Manual review processes require a regimented Quality Assurance Program to ensure that decisions are made consistently and to prevent any regulatory compliance issues. They are also an important component of developing automated lending process feedback loops. Quality Assurance managers should add value to the automated lending process by providing feedback about credit analyst training needs and trends in the populations being reviewed.
The lending industry continues to be challenging for loan marketers, but there are still profitable segments available to those who are prepared to work more closely with their internal credit policy and operations groups to put some of the art back into the science of lending. Progressive financial service companies who are prepared to take the extra steps required to identify and eliminate missed opportunities will not only survive, but will, in the long run, be rewarded with customers who are more profitable and loyal.
Bridgeforce, a women owned business, specializes in serving the needs of companies that either directly manage account receivables and financial transactions, or that aim to provide software and technology products for this market segment. Our services span across multiple industries, including: consumer lending, small business lending, auto lending, retail banking and brokerage, healthcare, insurance, telecommunications, utilities, and call centers.
Bridgeforce is a partner of ELSOS, for more information please contact us @ 760-438-1741 or firstname.lastname@example.org