In my first blog, we addressed the definition of auto decision and financial institution Culture. Now, let’s look at data analysis.
There are many pieces to data analysis, and I am not a statistician, so what I will address here are the basics.
You cannot auto decision day one!
- Many clients purchase a system like Bank2Business and want to auto decision day one. The financial institution has no experience with business scoring, but they want the efficiency immediately. I am often asked, “Why I can’t tell the bank which scores are good?” The answer to this is in my second point…
Auto decision is an automated replication of your underwriters’ decisions—that requires data.
- We can’t mimic your underwriters’ decisions if we don’t know what they are. Every bank has a different risk tolerance. A good score to one bank may be too risky for another. Until we can analyze that information, the gray area strategy and other factors cannot be set.
- The MINIMUM number of loan decisions to consider creating an auto decision gray area strategy is 300. That means 300 approvals and declines of new money. No renewals, no cash secured credits. New dollars. And that is a minimum
Auto decision is not just “score and go”.
- Auto decision is more than just score. It is recommended the financial institution look at other, predictive factors as well. An analysis of high and low side override reasons should be done to determine the credit policy messages needed to recreate the underwriters’ decisioning. These messages may be time in business, deposits levels, etc. This is the type of data that is gathered during the manual underwriting processes and used to create the auto decision parameters.