December 10, 2012

Step-By-Step: Less Risk, New Challenges

While smaller financial institutions may be able to pull off a big bang core replacement, larger ones with more complex systems usually opt for a phased approach.

Phased Approach
  1. Eliminates the risk of converting everything all at once
  2. Breaks the project into shorter, easier goals, allowing it to build momentum
  3. Can use lessons learned in the early stages to apply to later stages
  4. Allows for the time needed to replace complex and tangled legacy systems

Banks with more than $5 billion in assets tend to do a staged replacement, and those with under $5 billion usually do a rip and replace, says Anne Miela, VP of program and product management at Open Solutions, a core provider in Glastonbury, Conn. "The larger ones have a more spread-out branch network and their data centers can be more spread out, too," she says. "They could have their headquarters in Detroit, but their mortgage center is in Dallas."

Many banks of all sizes simply have too much complexity in their legacy systems to be able to pull off a big bang core replacement. "You have to consider the complexity of the current systems, the architectural complexity and the number of users, as well as the vintage and range of systems. Some places have 30- to 40-year-old systems in place. That makes it impossible to do a big bang," remarks Ashwin Goyal, VP of product management at Oracle Financial Services in Redwood Shores, Calif.

However, while the step-by-step approach reduces risk for these larger core replacements, it also presents new challenges, such as working with coexisting systems during the replacement and managing such a large organizational change over time.

Oracle recently released a core banking system targeting bigger banks looking to replace their aging legacy core systems. It's designed to help them do a step-by-step core replacement; each architectural component of the core system can be broken down and managed on its own without affecting the other parts of the system, Goyal says. This eliminates a great deal of risk in the replacement, which is part of the trade-off in taking a phased approach: There's less risk since the whole system isn't being replaced at once, but it takes more time and more money.

Oracle's system lets architectural components of the core moved to the new system function with the architectural components still on the old system during the transition. This helps address one of the chief challenges of a step-by-step core replacement, Goyal says: working with coexisting core systems during the replacement. He compares a phased core replacement to open heart surgery (a common analogy in IT), where the old and new veins (systems) are running at the same time during the surgery (upgrade).

It's critical for banks to plan how to manage coexisting systems during the testing phase, Goyal notes. He recommends that they avoid splitting the customer relationship between the old and new systems during the replacement. Doing so, he says, can lead to confusion for customers and front-end associates, costing time and money.

Although a gradual core replacement can be a drawn-out costly process, Bruce Livesay, CIO of First Horizon, the holding company of First Tennessee bank, managed to complete a step-by-step core replacement for the bank under budget and ahead of schedule a couple of years ago. The Memphis bank ($26 billion in assets) broke the project into manageable short-term goals and that helped the project gain momentum as those goals were met. If you do the replacement with incremental deliverables, "you can get some wind in your sails," he says. "Focus on shorter-duration goals and celebrate them when you've accomplished those goals," he says. "That helps in dealing with burnout."