Building a Bank for Today from Scratch
Ever since cash dispensers became automated teller machines (ATMs), pundits have regularly said that the large retail banks are going to have to change their core systems or they will lose their customers.
It started with the ATMs because those machines broke many banks’ assumptions that their customers transacted all their bank business in their own branch. It only got worse when electronic point-of-sale got added into the ATM architecture. Then, when mini-computers and PCs took over from mainframes in the early ’90s, the pundits said that now the banks would have to abandon their out-dated core systems for cheaper alternatives. Next up was the Internet, which made it even more obvious that bank customers didn’t want to go to the branch, but rather wanted the bank to come to them. This was surely going to kill off the mainframe core banking systems.
Finally, along came mobile. One banker at a large institution told me that it took ten years for the Internet bank to overtake the branch while the mobile bank did it in a year. This was confidently predicted to kill off the mainframes about five years ago.
So, I am not going to predict the death of the mainframe core banking system, since that would be heavily redundant. Rather, I am going to describe what a bank would look like if I were to design it now, working from a clean slate, as it were.
Good as Amazon
First up, my new bank would focus on providing high levels of customer care. It would be at least as good as Amazon. Every interaction with the bank would be remembered and used to add value to the customer’s financial transactions. All services would be available on all appropriate channels 24 hours a day, seven days a week. By contrast, today’s big banks are accounting focused; in fact, their systems are called “branch accounting systems.” This is why there is so little differentiation among the large banks.
Next, my bank would assume extremely low margins, so all services would be allocated their full costs, which would be charged to the customer at a small profit. No pretence of free banking.
Third, my bank would assume that it would get thousands more enquiries than actual financial transactions. Today’s banks are built for a “look-to-book” ratio of five to one; mine would assume 5,000 to one, as is the case for online travel and capital markets.
Today’s large core banking systems stop for the night. They close up after 6 p.m. and wait for the overnight batch accounting run to complete before starting the next morning at 8. My bank would do everything in real time. You should be able to take $100 out of the ATM at midnight and see the transaction on your mobile as you walk away.
My bank would also assume that the customer wants the bank to come to them for all transactions. This would be for paying in shops, paying online, paying friends and family, international payments, getting cash, getting foreign exchange and depositing funds. There would be no need ever to come to a branch to do business, but I would maintain some Main Street presence for customers who want to talk to someone face to face.
Not only is my bank real time, but there are no batches. All processes are same day and straight through. When I opened a mortgage in the U.S. with a large bank in 2000, it took six months for the bank to get the mortgage onto its own online banking system as all the different month-end batches were completed. This is too expensive, too error prone and too slow for the customer.
I wouldn’t build a data center for my bank. Instead, I would rent the hardware and software online using infrastructure as a service for my front office and software as a service for my core banking. This is what even large banks do today for credit card processing, but not for core banking. By sharing the core banking, which is a commodity to my customers, with hundreds of others, the cost can come down to less than 16 cents per account per month – about 100 times cheaper than an in-house mainframe. I also only have to pay for the processing I need by using the elastic provisioning of these services. This saves a fortune, particularly if I still run close-of-business interest capitalization.
Finally, I would treat my customers fairly and only make money by charging a fair margin on costs. I would never hold the customer’s money to take advantage of legacy payment schemes that take one, three or five days to clear funds.
Whether today’s large banks can do any of the above with their existing core systems I leave for you to decide.