One early learning I had in banking goes back nearly four decades, to the days of COBOL. In setting up an analytic tool for customers and credit cards, the team I was on decided to include a “bad check indicator.” In reviewing a customer credit card record, a simple counter provided an insight into how the household managed their finances. The good news was that bounced checks were an indicator of household financial management, the bad news was that we only used one position to count bad checks.
Who could have guessed that one digit from 0 to 9 would not be sufficient? In testing, the finding was that if the customer bounced 10 checks, the counter would show “0” rather than “10.” The issue was quickly fixed and carried out to three digits. An excellent boss explained the issue and it was a learning experience.
It turns out that once a household budget goes awry, that bad check counter will get a lot of work. In a bad month, one bounced check could cause several others to cascade, pushing the stressed household deeper into debt. So much for the young and naïve.