Thursday, June 19, 2008

Another benefit to new architectures

It's been said that bankers will take twice as long to adopt a new technology as just about any other industry. The cases used to illustrate that are generally the automated teller machine (about 10 years, though much of that was consumer resistance) and automated underwriting in the mortgage space (another 10 years). Sure, there are other industries that are also risk-averse that also move slowly into unproven technologies, but the old R&D/IT investment numbers that MBA chief economist Doug Duncan used to toss around stick in my mind. The mortgage business, he said, invests about one-tenth what most other industries do in technology research and development.

But why should they invest more? People that manage risk for a living don't like to take unnecessary risks. As long as there is an old DOS-based system cranking away reliably on the company's accounts, new technology could look like an unnecessary risk. But that may be changing.

The benefits of SOA-based architectures have generally been sold to the mortgage space bundled with MISMO data standards with the promise that they can reliably connect business partners in a fragmented industry without the high costs of system-to-system integration. But lenders are finding that there is another benefit to having loosely coupled systems that can easily be plugged together: making pilot programs easy.

Now, with the insertion of a single business rule, lenders can test out a new fraud detection system by simply telling their processing technology to send every loan that is, say, over $300,000 and not owner-occupied in a certain ZIP code to a new company for analysis. If there is no ROI, they can turn it back off again and go back to their old vendor. Easy. The same holds true for just about every other service the lender might need to close a loan.

Of course, it will take some expertise to build rules that provide good tests, but the point is lenders can do it now and it won't cost them big bucks to try out something new. How's that for mitigating risk?

No comments: