'Short-Term Value' Helps the Long HaulBy Anna Maria Virzi | Posted 2004-09-01 Email Print
WEBINAR: On-demand webcast
Next-Generation Applications Require the Power and Performance of Next-Generation Workstations REGISTER >
This CIO learned early on why it's important to break big projects into small piecesand move fast.
I learned how to deliver applications in 30- or 90-day cycles at Lehman Brothers in the early 1980s. This was long before the dot-com bust brought financial sanity back to planning projects, particularly those involving information systems.
In an environment where traders work by the minute, you discover pretty quickly that long-term initiatives don't survive. If you tackle a two-year infrastructure project, you must build a road map and deliver changes incrementally.
At Lehman, we wanted to drive greater value from computers.That included offloading expensive computing cycles from mainframes to local networks of servers and "client" machines on desktops. This was a multi-year effort across multiple product lines.
In the midst of that project, we met with the head of the medium-term notes desk to figure out what the technology staff could develop to enable traders and salespeople to do their jobs better. One idea: create a new system enabling us to use packaged reporting and analysis software to create high-quality portfolio statements for customers.
We examined our database of customer account information, including customers' bond holdings and their valuations, as well as transactions showing when bonds were bought or sold on their behalf. We then collected information about market trends, and used that data to create sophisticated reports for Lehman's best customers. We built an information system to make this happen in less than 90 days.
Why did this work? We reduced risk by creating a process of producing customer information without relying on the monolithic processes embedded in the mainframe application. This allowed us to eliminate changes in the mainframe and manage the solution at the client/server level.
This project was relatively easy to digest because we could focus the development on the task at hand, without disrupting processes embedded in legacy Cobol code. We were able to circumvent complex regression testing required if we had developed this application on the mainframe. Rather than being relegated to the bottom of a long development queue, the project took only two months from idea to deployment.
By breaking the changes into smaller chunks, we delivered short-term value by developing a product that could be used by our sales team to support their high-end customers and potentially upsell to them as well.
At American Express' global financial technology division in the late 1990s, when big projects were all the rage, I helped the company complete some major efforts pretty quickly. That included building applications enabling the company to consolidate financial transactions in three centers instead of more than 40 locations around the world. To make sure we could move quickly, we broke the project into smaller pieces. We started with the most important, accounts payable.
First, we defined our bill payment process and then developed best practices to distill the most effective elements of this process. We wanted to ensure that work done by the business operations staff could be passed off from one center to another around the clock in a standard fashion. Accounting clerks had to get secure access to accounts-payable data used to process payments and complete accounting journal entries. We created an application to address this in less than four months by partnering with the business side.
By following the mantra of delivering short-term value while engaging in very large projects, I've managed to lower business and technology risks while proving the value of these projects to the business. The complexity of the projects is reduced by breaking them into manageable chunks. Sometimes, you can't avoid the big projects that cross multiple product lines, such as building out a new business process for managing customer relationships. Even then, you can optimize the effort by breaking down deliverables into manageable chunks.
Charlie Schwab is CIO of Mercury Interactive. He previously worked at Lehman Brothers, JP Morgan, American Express and, yes, Charles Schwab & Co. "I'm not part of that Schwab family," he says. "We checked the lineage."