IT matters

Interesting discussion caused by an article in the WSJ about the decision of ASU's chief technology officer to forgo full testing in an ERP implementation. How did he do that? Well, he just took a page book out of all the Web 2.0 startups and released the software in beta (That's a polite way of saying the product/service is not fully finished/tested.).

So who tests the application? Well, the poor users of course. And this is what has caused a firestorm. There have been cases of employees not getting paid, or in some cases overpaid. The CTO's justification for not fully testing? He's apparently managed to save over half the original $70m estimated price tag and reduced the time taken to bring the application to its users. Very impressive. But is this a strategy that other CIOs should follow?

I think not. While releasing a photo sharing program or email software such as Gmail in Beta can be justfied, doing this for software that going to matierally impact people lives can not.

• How would you feel if you tried to book a ticket from Phoenix to Chicago but ended up getting a ticket from Phoenix to San Diego? Right.
• What if you tried to order a shipment of 1000 pens for your company using your procurement system, but instead got a shipment of 1m pens? Oops.
• What if your mortgage company's loan origination system mishandled your closing date so that you couldn't close on your home? Hmm.

What's really shocking is that one would skimp on testing. Of all the items that one can offshore, testing and documentation of software are probably the items that are the least risky to offshore. Why for say $2m, the CIO could have got a team of say 65-70 testers testing the crap out of the system for say 6 months. Surely a wise investment given the $30m price tag.

What do you think? When is not testing software justified?

Posted by Akash Bhatia at 05:29 PM | Permalink | Comments (3) | TrackBacks (0)

When I think of IBM, I think of hardware, software and consulting services - I don't think of it as a lender. Yes, that's exactly what IBM has got approval to do.

" IBM Lender Business Process Services, or LBPS, received clearance to provide mortgage origination services for federally insured Federal Housing Administration (FHA) loans.

When it announced the unit’s launch in March, IBM said the unit would fill a void in the lending space, particularly for small and midsize lenders.

The Charlotte-based unit will allow mortgage lenders to replace the fixed costs that are associated with typical loan fulfillment operations with a variable-cost framework. This in turn will free up lenders to provide better service and support to consumers, IBM says.

LBPS will offer a variety of lending services, including loan application, underwriting, processing, vendor management, document preparations, and loan closing."

While this may not seem like a significant announcement, I think there's more here than meets the eye. This is really IBM introducing LaaS (Lending as a Service) to the market.

Lenders typically spend on the order of hundreds of millions of dollars building and maintaining custom loan origination systems. Such projects are a CIO's worst nightmare - budget overruns and long delays seem to be common. IBM's offering of a no money down, pay as go model seems to be a god send. But this may not fit all lenders equally well. 

I don't see large lenders (Countrywide, Wells Fargo, Bank of America etc) using this. Large lenders typically have a great deal of blood, sweat and money invested in their origination and fulfillment systems. It's not an easy investment to walk away from, especially since each lender will claim that their processes are different and hence they need a custom system. Moreover, it's highly unlikely that they'll risk exposing their decisioning/pricing rules - really their secret sauce - to a competing lender.

However, I think IBM is bang on in offering this to smaller lenders. First, they don't have to spend a fortune standing up yet another origination system. Second, the variable per-loan cost arrangement will allow companies to better weather markets turns.

As with any SaaS offering, it will be interesting to see if they'll be able to customize the experience, processes and rules for each client without incurring exponentially rising costs. While its fairly obvious how they're planning to use their recent acquisitions of FileNet (BPM) and Palisades (Origination/Fullfilment) in this offering, it's less clear how's they're planning to offer a custom user experience for each of their clients. Adding a dash of Bowstreet's Factory (User Experience Customization) to this recipe may make this a potent offering.

Posted by Akash Bhatia at 04:05 AM | Permalink | Comments (0) | TrackBacks (0)