Q&A - Allan Woods, vice chairman and chief information officer at Mellon Financial

WS&T: How is your technology budget defined?

WOODS: We've been fairly consistent. About 12 percent of our revenue is technology expenses. When you go back over past few years, that's where it has been and that is how we anticipate it being next year as well. Half (of that budget) is infrastructure - big mainframes and networks. The other half goes to application development.

About 70 percent of total-technology spend supports our corporate and institutional services sector, the sector that houses our large processing business. It makes sense that they take the lion's share of our technology resources. The remaining 30 percent goes to our other major sector, our asset-management sector including the private-wealth-management business, Dreyfus, and all of our investment subsidiaries.

WS&T: What are some large projects that Mellon has undertaken in the past year?

WOODS: In the past year, we've developed interesting technology around straight-through processing which involved imaging, character-recognition and workflow technologies. A good example is that we have a business in our cash-management business called a lockbox business. This is a business that gets some 30 million items to be processed every month. With imaging and character recognition, 70 percent of those transactions go through without any manual intervention at all. It's also allowed us to do account-receivable conversions. We can take a company's manual paper-based check payments and convert them into automated clearing-house items. Benefits for the customer are that it gets them money faster, and, as their transactions and go through various clearing houses around the United States, if there are any rejects, they get handled first. So it clears up any funds that are in suspense very quickly. It's a very popular feature of the cash-management business.

This year we've also spent a fair amount of money with 25,000 desktops. We've invested in technology that allows us to put them all on the same operating system, to make sure that they can all be managed from central locations, so you no longer have to dispatch people on the sneaker patrol to go to your desktop to do things. (This includes) everything from virus definitions to patches to monitoring the help on the desktop. It's all done from a central location. That's reduced our cost to maintain these things dramatically.

WS&T: What is your project-prioritization process?

WOODS: From a technology standpoint, the infrastructure is truly centralized. Application development is centralized, but each line of business decides how much they want to spend and how (they want to spend it). We are not making the decisions we shouldn't be making. (Lines of business) are making those decisions.

WS&T: How do the lines of business understand how much to budget for technology?

WOODS: Each one of our major lines of business has a chief information officer that we assign. That individual is the focal point of communication. He sits on the senior-management group of that business. When the project is first discussed, based on a high-level design, we'll give the line of business an estimate. But we can't commit until we do detailed specs and design. The line of business is prepared to make a go/no-go decision once the detail is made.

We're a technology company in disguise. We sell our technology, so most of these individuals are looking at their businesses, understanding their needs to invest in technology. They are all aggressive users of technology so they'll estimate, based on their operating plan, how much they can carve out on technology.

WS&T: What applications are more suited for buying from a third party? What is more suited for building?

WOODS: Thematically, building is the choice of last resort as far as I'm concerned. I would much rather buy than build. The only reason I'd build would be if there were no industrial-strength solution. One of the big functions in asset-servicing business was, how do you handle corporate actions? That is a very complicated process. We were looking for a system to do (corporate actions) and we couldn't find one, so we entered in multiyear project to build our own. If there were one on the market, I would have bought it in a minute.

WS&T: Where are largest costs in building applications?

WOODS: The most expensive part of a project is the testing on it. Coding is next, and the whole process of specification, requirements, and design, and then training. Then the project is delivered and then there is maintenance. That is a separate discipline.

WS&T: Are there any timeline restrictions on building?

WOODS: Any project that goes over a year is a project that is very difficult to manage the returns on. If there is a project that is long, we look for value as soon as possible.

WS&T: Is software quality a concern in the industry? How can you ensure that Mellon is both investing in and developing sound applications?

WOODS: We've been investing over the past four years in Capability Maturity Model (CMM). It's a model for developing application technology and software that was built in Carnegie Mellon University Software Engineering Institute. It's a process " the most renown developers in the world subscribe to this. It's kind of like the martial-arts system. You get basic ratings zero to five. We convinced ourselves that, given the fact that we really sell our technology, we've got to have this as a competency. So our objective was to get all of our application-development units up to level three. By end of this year they will all be at level three. The large ones are already there.

WS&T: When do you expect to move to level four?

WOODS: We're not sure we understand the business value of going beyond that. We think it's a point of differentiation for us. Only six percent of involved companies are financial institutions. It's really hard to do, it's a major sea change. It doesn't just involve technologists but also lines of business.

WS&T: What is your perspective on outsourcing versus staying in house?

WOODS: We do a number of different kinds of outsourcing. We'll outsource certain functions to organizations that have applications that we consider best-in-class applications. We outsource our accounting for our private-wealth management (business). There are firms that invest in these applications consistently so it's something that we don't think makes any sense to do on a proprietary nature. Our objective has been to outsource 25 percent of our development functions to offshore by the end of next year, and we're almost halfway there.

WS&T: How do you establish possible outsourcing projects?

WOODS: The key with what we outsource and what we don't is first to make sure there are no contractual restrictions. We also need to make sure that the outsourcing firm can develop technical expertise to do it, and usually that's not the problem. We have found very few barriers to offshore outsourcing of application development. Part of that would be new applications that we're developing, and we choose those based on how much hands-on knowledge do you have to have about what you are developing. The other piece is the maintenance functions. You try not to outsource all of your expertise on any particular function or application because you want to make sure you always have contingency plans.

WS&T: Would you ever consider outsourcing your infrastructure?

WOODS: Over the past five years we've had three major infrastructure engagements done. That's where we appoint a third party to look at our go-forward model and an outsourcer's model, and determine if there is economic incentive for us to outsource our infrastructure. It's uncanny but all three of the studies came back the same: that there really aren't sufficient economic incentives for us to outsource our infrastructure. But we'll look at in again in a few years.

WS&T: How has Mellon been concerned with rising security threats to financial institutions?

WOODS: This is pretty subjective. We try to find what we consider best practices around safety and soundness. Given the information we have, and our position in industry, we feel that we need to emulate best practices. Given all of the issues with the internet, that is a place we need to spend money. You can't map out return on investment, but good judgment dictates that (security) is a sound investment. When we go to our lines of business and chairman and declare that something is a safety-soundness investment, there is very little pushback.

We've increased the investment in corporate-information security. We've increased the number of threat and vulnerability tests that we have done by outside concerns. Every one of our Internet-facing servers goes through the threat and vulnerability tests at least twice a year.

WS&T: What are some emerging technologies that might get some play time in your firm in the next year?

WOODS: We've just built a huge new Mellon Financial center in London. That's an excellent environment to put in Voice Over Internet Protocol. We haven't moved in yet so we haven't gotten some of the experience in actually using it.

We have a couple of pilots with Linux. The challenge is if you wanted to deploy it more broadly, there's an expense of getting from where you are onto that platform. As much as I'd like to wave the magic wand and put more things (on Linux), we must be careful about that investment.

We're devoting some time to collaborative technologies like the Internet for shared meetings. It has significant advantages and it's also become an interesting customer-relationship tool as well.

We expect significant application impact when that critical mass has been achieved (in wireless and mobile technologies). Critical mass is the key. We tend not to wait until something is proven to get involved. By then you are so far behind on the learning curve that it takes too long to get up to speed. We pick things we think will be technology winners.

WS&T: What major IT projects are on the table for next year?

WOODS: We have some major new business wins that require technology to bring them on. That's going to take quite a bit of technology expense for integration of the new pieces of business and also product enhancements that new businesses would require.

We want to build out more of our asset management back-office platform. We do outsourcing for asset-management back offices and next year we'll have significant enhancements for that.