Overcoming the Fear Factor: Migrating Core Banking Systems
There are many drivers in today’s banking environment that are stretching core systems to their limits. Channel and product proliferation have created increasing complexity. Rolling out new products now requires that many channels be supported, dramatically increasing the time from concept to launch. There has been focus on customer centricity driving cross-selling and relationship pricing, and legacy systems have been scrambling to catch up. Customers are now demanding real-time information like never before, via the Internet, IVR, ATM, and branch. These channels all need to provide the same answer (and the correct answer) to the customer. Data requirements of Basel II are much more easily delivered with real-time relational databases than they are with batch systems running flat file databases.
The number of connections that must be maintained with legacy systems to the front end has grown consistently over the past 30 years as shown in the figure below. This has created huge demands on IT departments because they must integrate and test new products against a myriad of channels, slowing down product innovation.
Yet the core systems remain unchanged. Each time a new product is introduced, new interfaces need to be created to each of the channels. Each line in the figure above represents custom code that must be written and maintained. Legacy systems were not designed to interact with a multitude of front ends or a multitude of other systems. Modern core systems support a service-oriented architecture (SOA) and some sort of common framework of messages such as Interactive Financial Exchange (IFX). This makes interfacing to various channels much more manageable. Additionally many of these channels are expected to be real time. A customer doesn’t want to wait a day to see that his direct deposit came in or that his debit card has brought his balance below zero.
Banks have learned that they can improve their overall profitability through a thorough understanding of their customers. If the architecture of the systems is based on product silos, it is more difficult to understand customer behavior or profitability. Research shows that the greater the number of products the customer has with the bank, the more profitable and the more loyal that customer becomes. Bankers across the world have understood this message and have major initiatives for cross-selling into the existing base to increase the number of products per customer, which will drive both profits and loyalty.
In order to support these initiatives banks require a customer-centric banking model, where customer-facing personnel can see the entire picture of the customer’s dealings with the bank and act as an advisor, providing consultative selling. This is a challenge for the existing core infrastructure. Legacy systems are product based or account number based, and in most cases the customer is a separate entity in each silo. What this means is that if you hold a checking account, credit card and mortgage at the same bank, it is quite likely that the bank will need to change your address three times in three systems.
Basel II is imposing new risk measurement on international banks, and many are finding that they are not able to extract the information they need to meet regulatory requirements. Modern core systems designed on top of relational databases allow for the extraction of data in a batch or non-batch environment. Just as there is a huge data integration challenge simply to identify customers at a bank, there is a similar integration challenge to meet Basel II requirements. This again puts strains on the legacy systems that were optimized for transaction processing rather than real-time data extraction and analysis.
To bring some real world perspective into this article, one should examine Webster Bank, a US$20bn bank based in Connecticut. This bank accomplished the feat of migrating 20 major systems over a 15 month period from system selection including core DDA, teller, platform, commercial lending, and ATM.
Webster Bank grew from being a US$12bn bank to a US$17.8bn bank in just two years. It currently ranks 59th in assets among US banks. Zeynep Fredrick, the CIO, was brought on board specifically to address the question of whether the existing core system needed to be replaced. Fredrick was faced with the task of creating the IT infrastructure for the bank to grow to double or triple its size.
The bank was motivated to move core systems because it had converted from a thrift to a commercial bank and found itself limited in its ability to present a broader product offering to commercial customers with its existing core solution, Fidelity (Aurum) Miser. While Miser was very good at thrift processing, with real time straight-through processing and an integrated suite of products, the platform was missing cash management and commercial lending functionality.
There are a number of surprises about this core migration. First, the bank chose a ‘big bang’ migration, replacing nearly all of the bank’s IT systems at one time rather than using a phased approach that is more widely practiced at large financial institutions. While many CIOs would consider this too risky to attempt, Webster Bank found it to be the best solution for the bank and succeeded.
Secondly, Webster also chose to outsource the majority of its new core systems rather than keep them in-house. “People who knew me were surprised,” says Fredrick, “but I needed to remember that this conversion was for Webster Bank, not for IT.” Given that the bank was running Unisys mainframes and the new core was running on IBM mainframes, there would have been large conversion, training, and hiring costs associated with bringing the mainframes in-house. The cost of building up the IBM mainframe infrastructure was huge with mainframe hardware, operating system installation, and direct access storage device (DASD) management. The conversion costs doubled for an in-house deployment at Webster compared to outsourcing. While Fredrick acknowledged that she preferred an in-house deployment, she ultimately decided that the benefit was not worth the cost. She was also persuaded to do this because other large banks had outsourced core functionality.
The drivers to change core systems were:
In response to these challenges, Webster Bank undertook a massive core conversion. The speed with which this migration was accomplished is also unprecedented for a bank of this size. They completed the entire process in two year, as shown in the figure below.
With focused and intense effort, even a big bang core migration is possible within a relatively short timeframe. The bank is now positioned to move forward on a new customer-centric, business-friendly platform.