Kevin Murray, CIO of domestic brokerage and personal lines forAmerican International Group (AIG), had a mainframe full of fat-clientlegacy applications. He just trashed it in favor of newly writtenthin-client Java and XML applications.
Dan Roberts, CIO of the PMI Group, a mortgage insurance company, is inthe middle of Web-enabling his back-office legacy systems, a projecthe anticipates will take up to three years. However, he says it's"absolutely necessary" if his company is going to keep up with productdevelopment and customer demands.
Maria Fitzpatrick, CIO of PacifiCare Health Systems, just decided toget away from the multiple OpenVMS and Unix programs scattered acrossher company's business units, and upgrade to a single Web-enabledplatform. The project is part of a major effort to redesignPacifiCare's corporate strategy by unifying business processes acrossall units.
The list, long now, will get longer. Every day thousands ofenterprises rely on decades-old applications written in obsoleteprogramming languages that, along with the systems on which theyreside, are no longer supported by the application's creators - whoeverthey were and wherever they may be.
So why don't CIOs just pull the plug on these ancientapplications?
"If it was that easy to get off these systems, most CIOs would havedone it already," says Dale Vecchio, research director of applicationdevelopment for Stamford, Conn.-based Gartner.
"The only way I'll address renovating my legacy systems is if theystop enabling the business," says John White, vice president of IT atParker Hannifin, a Cleveland-based manufacturer. "My barometer is theROI," he says.
While White's rule of thumb is widely held, more and more businessexecutives are coming to grips with the fact that the Internet hasplaced demands on companies and computing systems - such as real-timeorder processing and managing high-bandwidth demands - that most legacyapplications just can't handle. CIOs know this better than anyoneelse. But right now would be the worst time for an organization tobegin a major infrastructure face-lift, wouldn't it? The economy is inclampdown mode. Enterprises are struggling to find cash for the mostbasic projects. Only the nuttiest CIO would argue for spending moneyon infrastructure, and only the most irresponsible CEO would approvethe expenditure. Right?
Actually, not right. In fact, spending money to keep legacyapplications going is a mistake. Assigning personnel to keep legacyapplications running is a big mistake. Making business plans based onlegacy applications is an enormous mistake.
Despite all apparent evidence to the contrary, right now is the idealtime to either pull the plug (which would be ideal) or overhaul legacyapplications. And a whole roster of major American enterprises areseizing this opportunity to get a jump on their competitors bymodernizing (connecting their legacy system to a Web front end) ormigrating their legacy systems to new thin-client-based Web systems.
And they're doing it right now.
The Legacy Albatross
Legacy applications underlie almost every enterprise. In general,these applications are stable but inflexible, expensive and difficultto maintain. How expensive? According to Gartner, between 60 percentand 80 percent of an average company's IT budget is spent onmaintaining existing mainframe systems and the applications that runon them. Maintenance of legacy software is complicated by the factthat the number of programmers who know how to handle former standardssuch as Cobol is shrinking with each passing year.
Though not every legacy application needs a complete overhaul, theharsh reality is that all legacy systems at least must be Web-enabledif a business is to grow and remain competitive, says PacifiCare'sFitzpatrick, senior vice president and CIO of the $12 billion SantaAna, Calif.-based company. Fitzpatrick plans to migrate from threedisparate legacy software systems to a single, Web-enabled platform,which she hasn't yet chosen.
"Right now, cost efficiency is key," she says. "Maintaining and tryingto integrate multiple software platforms becomes expensive over time.There's a great business need for companies to strive to provideincreased service levels to customers. Web capabilities arethe path."
The project's expected ROI stems in part from the projected decreasein maintenance costs of the new system compared with the legacysystem, she says. Also, by linking to the Web, Fitzpatrick can put hercompany in closer contact with its constituents - doctors, hospitalsand employers - and decrease administrative costs.
"The most important justification for beginning this kind of projectis the business need," says Bruce Fadem, vice president and CIO ofAmerican Home Products, a $13.3 billion global pharmaceutical companybased in Madison, N.J.
Because legacy applications are so tightly tied to the way a companyfunctions, the health and flexibility of those applications directlyaffect the business' ability to grow.
"The time is right for modernization because the attempt to marry theworlds of legacy and e-business has reached the point of pain," saysTyler McDaniel, an analyst with the Hurwitz Group, an IT consultancyin Framingham, Mass. "If you wait to modernize, you'll continue tosuffer consequences such as errors, delays in processing andfulfillment lags. These are things that will drive your company out ofbusiness."
You Are Your Legacy System
Before beginning any kind of modernization or migration project,business and IT executives must understand exactly what their legacysystems do and whether the way the systems work actually reflects thebusiness strategy, McDaniel says.
"It's a huge mistake to jump into migration on a per-project basisjust to see where it leads," he says. "Develop a vision of where youwant your architecture and infrastructure to be in two, four, six andeight years. Assess what skills are available and know what yourinvestments are right now in development versus maintaining legacyapplications. Does the existing system map the key business processesthat are critical to success? That's where you must start."
David R. Guzman, senior vice president and CIO of Glenallen, Va.-basedmaker of medical and surgical supplies Owens & Minor, did just that.His legacy systems were configured to assume that Owens & Minor, with$3.5 billion in revenues, owned all the supplies it shipped tocustomers, but the company was evolving toward a third-party logisticsdistribution model. What Guzman needed was a system that allowed thecompany flexibility in distribution methods. This provided thebusiness rationale for migrating to a Web-based platform.
For Roberts, CIO of the $763 million San Francisco-based PMI Group,his legacy system's ability to support customer demands was"questionable." The back-office systems, which included claimspayment, billing and policy maintenance, were 12 to 15 years old.According to a 1996 study commissioned by PMI (one year before Robertsarrived), the cost of modernizing the company's legacy systems,estimated to be about $20 million, wasless than the estimated cost of the potential service-related failuresthat could result if the systems were left in place.
At that time, PMI staff in both the IT department and the policyservicing department were spending hours cross-referencing customerdata between the legacy systems and the separate policy systems. Asthe two systems didn't talk to each other, there was always the riskthat someone might miss something during the manual check.
"It was only a matter of time," Roberts says. "Everyone knew this wasa long-term problem and that our business was getting more, not less,complex."
After careful analysis, Roberts determined that the databaseunderlying his systems wasn't inherently flawed. He stuck with theAS/400 platform and is rebuilding his back-office applications in Javaand RPG, an AS/400-specific programming language. He's looking atseveral legacy migration tools that could handle translating the dataanalytics processes hidden within his legacy transaction processingsystem, but he has not decided yet on a particular tool.
PMI's executive board members knew the legacy system was a problem,but they delayed the migration project in order to focus on gettingthe company's e-commerce strategy in place. Once that was complete inlate 1999, Roberts and Kathy Schroeder, vice president of policymanagement systems and the project's business sponsor, got to workconvincing the board that the time was right to tackle the migration.And one of the reasons the time was right, they said, was that thecost of the project, estimated at $20 million in 1996, had come downin 1999 to an estimated $12 million.
"The board knew that the business had changed and that the legacysystem wasn't built to deal with the kind of claims processing andtransaction products we had," Roberts says. "They knew that if wecreated a new product, the old system would limit our ability to rollit out. We had to make it very clear that waiting any longer couldreally hold us back."
The migration was approved, the money found. Roberts went to workpulling apart his legacy applications in October 2000. He plans tofinish by the end of 2003.
Selling Your CEO on Legacy Modernization
American Home Products' legacy system, according to its CIO, was readyfor the junk pile. "We were dealing with a system that was weak infunction and antiquated in platform, and to redefine our businessprocesses we had to reconfigure the system. It was that simple," saysFadem.
When multiple, disparate systems begin slowing a company's growth,it's time to take action, no matter the state of the economy. Butfirst the executive board has to give the thumbs-up. And when timesare tough, selling infrastructure projects can beproblematic.
To successfully sell a large-scale migration project to executiveleaders, CIOs must present decisive evidence that the project willsave money and strengthen the business.
"You have to fully understand the benefits of migrating, and discussthose benefits from a business perspective by showing how the projectties in to the company's strategy," says Al Biland, CIO of Snap-On, aKenosha, Wis.-based power tool and equipment manufacturer with $2.1billion in revenues. "That can be by making the company moreoperationally fit, by cutting costs or by generating profitablegrowth."
When Murray became CIO of New York City-based AIG, no one knew howmuch money the $46 billion financial services company was spending onmaintaining its legacy systems. He organized a total cost of ownershipstudy that served as a benchmark for the cost of maintenance andillustrated how much AIG could save in terms of money and efficiencyif the company migrated to a Java-based system. The figure Murray cameup with got the attention of his executive board. (Murray declined toshare that figure with CIO.) "I told them that we could be saving 30percent of what we were spending on maintenance if we moved frommainframe to thin client," he says. "Their jaws hit the floor. Theysaid, 'It costs us that much?"
Murray presented his board with the ROI of migrating over a five-yearperiod and showed how the project would increase employees' efficiencyby reducing the hours spent on manual processes and maintenance. Thestudy indicated that a migration would drastically improve AIG's speedto market and its customer service. That, Murray says, made thebiggest impression. "The executive team liked hearing how the projectwould improve the business," he says. "And they gave us the money weneeded."
In a cost-cutting environment, the theory that migrating now willprovide a competitive advantage later needs to be advanced withcaution, says Wayne Kernochan, managing vice president of platforminfrastructure at the Aberdeen Group, a consultancy inBoston.
"CEOs and CFOs have to look at the issue in both the short and longterm," Kernochan explains. In order for them to believe that their CIOunderstands their position, the CIO must acknowledge that the ROI ofinfrastructure modernization will not immediately be realized even ashe demonstrates that spending now will reduce maintenance costs. Atthe same time, the CIO can argue that "once we're out of the downturn,the advantages of revamping business practices now will become veryclear," says Kernochan.
Plus, the pain of migration, in both cash and time, is not what itonce was.
So You Think You Know Migration
Eight or nine years ago, there were few options for migrating orrenovating applications. CIOs could either re-build or replacesoftware systems, and either choice was extremely time-consuming andexpensive.
In 1992, Snap-On's Biland decided to get rid of the Cobol-based,homegrown IBM mainframe applications that his company relied on andreplace them with a Baan platform. "We had green-screen terminals andnot a lot of functionality," Biland recalls. "The legacy applicationsdidn't give us the level of detail around inventory and transactiondata that we needed."
Six years and millions of dollars later, the new platform was inplace.
Today, businesses have more options than replace or rebuild. There aretools that can do anything, from delving into legacy data, pluckingout relevant business rules and rewriting them in Java or XML, toattaching a Web front end on to an intact legacy database, andeverything in between. The tools are fast and relatively cheap,particularly compared with the cost of migrating an enterprise to anERP platform.
Owens & Minor's Guzman, for example, had no desire to go through anERP implementation. "ERP projects take too long, they generally don'tturn out well, and they cost too much," he says. But something neededto be done with the cumbersome, multilayered, 15-year-old applicationsthat contained the contracts and pricing systems for the company. Toupgrade, Guzman couldn't go to the originalvendor - KnowledgeWare - because it was out of business. And Owens &Minor's heavily customized applications ran OS/2, which IBM no longersupports.
"Legacy applications are a cancerous problem that needs to beexcised," Guzman says dramatically. "It's not a choice. If yourbusiness is changing, your systems have to change. We needed theability to do things like multicurrency transactions over the Web, andthe old system was completely inflexible."
Still, the old contracts and pricing system contained extremelycomplex business rules and mathematical computations that were vitalto Owens & Minor's dynamic pricing methods. The legacy system set aseparate, unique price for every product for every single Owens &Minor client, taking into account literally dozens of factors. Guzmandidn't even want to think about losing that functionality. So he tooka gamble on a tool by Relativity Technologies that promised to uncoverthe vital business processes buried in the old system, turn them intoWeb-ready components and translate the Cobol and CICS intoJava.
Guzman tested the tool on three master files containing EDI maps andcustomer information. Within six months, the test project wascomplete.
"It easily would have taken us years to modernize via hand-coding," hesays. "And it easily would have cost us tens of millions of dollars togo with SAP or Oracle. So far we've spent about $1 million on thisproject."
Guzman is taking the transformation one step and one application at atime. "We should complete the whole thing within 18 months and forless than $5 million," he says.
No Time Like Right Now
Modernizing legacy applications is a huge task, and it has both risksand rewards. It's easy to rely on what seem like stable systems andhope they'll sustain the business through a recession. But thesesystems are the technical avatar of a company's business strategy, andif they aren't updated, the business will begin tripping over its ownvirtual feet. Times are tough, but breaking a few proverbial eggs nowcould put your company 10 steps ahead of the competition. And that'spriceless.