Decision Engine Optimization. Presented By: Melton Knight, Senior Consultant at Experian; Credit Union Panel

Transcription

1 Decision Engine Optimization Presented By: Melton Knight, Senior Consultant at Experian; Credit Union Panel

2 Agenda Purpose of session Organization of this session Intended outcome

3 Melton Knight, Senior Business Consultant

4 Decision Engine Approach Why should you focus on your Decision Engine now? What are the things you need to consider? What is the value of this initiative?

5 Why should you focus on your Decision Engine now? As more Unbanked and Underbanked consumers enter the Marketplace, we need to begin to incorporate more non traditional credit information into the underwriting process. The following information may be accessed via your decision engine: Rental payments Public record data Thin file/no file scoring models

6 Why should you focus on your Decision Engine now? Your existing underwriting criteria needs to be constantly evaluated to determine where missed opportunities or inconsistencies may exist. By looking closely at the decision elements, you are able to identify which elements are directly impacting the final result. Where necessary, you can work with Vendors to make the needed adjustments.

7 What are the things you need to consider?

8 What is the value of this initiative? The performance of previously declined applicants can be re-evaluated to determine which could/should be reconsidered for approval by swapping-in. Likewise, those previously approved applicants that are under-performing may be swapped-out. By making adjustments to the existing criteria, improvements can be seen in your overall approval rates and BAD rates.

9 What is the value of this initiative? Based on the historical performance of past applicants, a confidence range can be created to measure the ability to accurately predict those that actually become good or bad. In certain cases where the confidence value is very high (90% or higher), there is a greater certainty of the outcome. These applicants can be auto- decisioned.

10 Panel Discussion Participants: John Listak, Alliant Credit Union; Danielle Bridges, Vantage West Credit Union and Cary Shumway, Arizona Federal Credit Union

11 Alliant Credit Union Presented by: John Listak, Project Manager LOS: Self Created

12 Alliant Credit Union Chicago, IL $9.3 Billion in Assets Symitar and Self Created LOS 75% Loan to Share Target 50% Instant Decisions Lending Focus: Direct: Auto, Visa Indirect: Auto, RV

13 Alliant Credit Union: Current State About 20-25% of loans are auto-approved No breakdown available by loan type One path per product Over 40 rules for New/Used Auto loans Very difficult to manage/make changes Auto declines rules: Age (under 18), Fraud, charge-off, delinquency with ACU Credit score ACU aggregate

14 Alliant Credit Union: Current State

15 Alliant Credit Union: Future State Ideally around 40% of loans are auto-approved Multiple paths per product One set of rules for current members One set of rules for non members Second chance for LTV, DTI, UDTI, PTI, and Based off of 26 attributes Focus on credit score Not all used for every loan type Auto declines rules: Age (under 18), Fraud, charge-off, delinquency with ACU Credit score (varies per product) ACU aggregate Availability of automated testing Ability to test rules before making updates

16 Alliant Credit Union: Future State

17 Vantage West Presented by: Danielle Bridges, VP of Consumer Lending LOS: Lending 360 by CU Direct

18 Vantage West Credit Union Tucson, AZ $1.6 Billion in Assets Symitar and Lending 360 Aggressive Out-Bound Sales Effort 105% Loan-To- Share Ratio Major Player in AZ Indirect Lending

19 Vantage West Decision Engine Features Over 1,200 variables Flexible Custom Rules Intuitive Easy Interface Control to make changes 3D Matrix Cross-Sell Module Challenges Pass/Fail Environment Not programmed for Premier attributes Not designed for multiple users/managers Does not pull info from core

20 Vantage West Current State

21 Vantage West Auto Decision Potential Unsecured About one-quarter of all unsecured loan/lines with FICOs greater than 687 are declined today. We approve about 55.5% of unsecured loan/line applicants and about 88% of these book. By implementing a recommended auto approval parameters, we could have auto approved 65% of the loan applications and maintained the same bad rate as we had on the 55.5% that we approved. Secured We currently approve about 52% of auto loan applicants and about 60% of these result in a booked loan. By implementing a recommended auto approval parameters, we could have auto approved 58% of applications, while keeping the bad rate at the present level of 2.75%

22 Vantage West Current State Lending 360 Decision Engine 15% Loans are auto approved (Secured and Unsecured Loans) Minimum 640 FICO, Max 45% Debt Ratio No Auto Declines Built in alerts for referral reason loans Test Environment for Rules Decision Challenger Flexible Decisioning for Unsecured Automatically includes stipulations (e.g. proof of income) System can be set up to counter offer

23 Vantage West Credit Union - Outlook

24 Arizona Federal Presented by: Cary Shumway, VP of Lending LOS: Meridian Link

25 Arizona Federal Phoenix, AZ $1.5 Billion in Assets Symitar and Meridian Link Prev. LOS: Datasafe with no decision engine Target 20% instant decisions Lending Focus: Direct, Equity and Credit Card

26 Results Overall:

27 Results by Product:

28 Recent Changes: Reject inference engagement Fast Start Cut off (640 to 620) Number of active trades (3 to 2) Length of credit history (36 to 24) Minimum number of trade lines (5 to 4) Process: Mobile loan application, automatic income waiver,

29 Reporting and Monitoring: Weekly sampling of decision engine approvals Trends Automated reporting through ARCU

30 Reporting and Monitoring:

31 Reporting and Monitoring:

32 Arizona Federal - Future Continue to capitalize on mobile technology CUneXus Decisioning as a Service (DaaS)

33 Questions

34 Contact Information: Melton Knight: John Listak: Danielle Bridges: Cary Shumway: