How MetLife Met the Integration Challenge

The austere bell-and-clock tower in which MetLife makes its home in Manhattan has been dwarfed by other skyscrapers since its completion in 1909, but it remains a monument in the Madison Square Park area. It looms as a symbol of the nation’s second largest insurer, with $2.1 trillion worth of insurance in force.

Last year, the MetLife companies served 9 million U.S. households, 4.1 million customers abroad, and 64,000 companies and institutions. It has 46,154 employees, and last year it amassed $32.5 billion in operating revenue. On the technology front, the company has five CIOs, one CTO and an executive vice president of technology who oversees them all. During its 134-year history, CIO-100 honoree MetLife has grown not only in size but also in complexity, becoming so broad that Chuck Johnston, vice president of insurance information strategies at Stamford, Conn.-based Meta Group, describes it as the “GE of the insurance industry.”

Much of MetLife’s scale is the result of its aggressive acquisition strategy — it has bought everything from billion-dollar enterprises to distribution channels across all its lines of business. Because of its rapid expansion, MetLife was saddled with a plethora of disparate systems and processes — in addition to its own legacy systems, some already decades old. In 1998, a new CEO, Robert H. Benmosche, came on board looking to transform the company’s reputation as a staid insurance company to that of a nimble, full-service financial services firm. To that end, he had MetLife look at its millions of customers from an enterprisewide perspective and start taking advantage of its enormity to cut expenses. At the same time, MetLife executives decided to take the company public, meaning it needed to respond to public-reporting requirements dictated by the Securities and Exchange Commission. All these changes meant one thing for MetLife’s technology team — it was time to get integrated.

Daniel J. Cavanagh knew big changes were ahead when he took over as executive vice president in charge of operations and information technology in March 1999. Appointed by Benmosche, he was given a $990 million budget. “We knew we were going public in a year. And we knew we were doing it very rapidly,” Cavanagh says. “So no one was taken by surprise.”

The message from “the chairman,” as most of the IT execs call Benmosche, was that the IT staffs supporting MetLife’s business units needed to rein in the subsidiaries so that the huge insurer could start reaping economies of scale. “When MetLife started testing its value proposition with the Street, the big sentiment was that it’s time to harvest some of that scale you’ve got,” says Tony Candito, senior vice president and CIO of MetLife’s individual business unit. “At that point, everything changed.”

Cavanagh began by “integrating” his five business unit CIOs and his CTO into an IT governance board. They meet monthly to discuss IT and integration strategy and specific projects, both enterprisewide and within each business unit, and they’ve had more than enough to fill their agenda.

Follow the Financials

The first integration task for MetLife was getting its financial house in order. That fell to Peggy Fechtmann, senior vice president, CIO of corporate systems and chief e-business officer. Even after the IPO in April 2000, its financial processes and systems remained siloed. Different subsidiaries sent financial information to MetLife’s comptroller in Tampa, Fla., in different ways — by FedEx, mail or e-mail — making it largely a manual job to close the books each quarter. “It’s comical in hindsight, but it certainly wasn’t comical then,” Fechtmann says.

Her first step toward standardizing and integrating financial systems was to form a steering committee of IT executives and subsidiary business partners to discuss the financial processes of MetLife and its subsidiaries. They knew they needed general ledger, accounts payable and asset management capabilities, and they planned to go with packaged software. “We have a lot less people doing hard-core coding, and a lot more people doing package implementation,” says Fechtmann. “We’re more systems integrators than anything.” They looked at Oracle, PeopleSoft and SAP before choosing PeopleSoft. “There was an exhaustive analysis, but there was a slight bias toward PeopleSoft because Met had implemented PeopleSoft for HR in 1995,” explains Fechtmann. MetLife committed $80 million over two years for PeopleSoft’s Enterprise Profitability Management (EPM) package.

Next, Fechtmann created a rollout plan. First was the MetLife company, or Mother Met as Fechtmann calls it, which was converted in April. Several smaller subsidiaries moved over in May, and several midsize subsidiaries, such as MetLife Securities, in June. MetLife Auto and Home is slated for the third quarter. After that come the larger acquisitions — New England Financial and General American Financial — in the fourth quarter, where there is sure to be a “fair amount of pain,” says Fechtmann.

Fechtmann and the committee allow no more than 5 percent customization of code — a conservative choice based on a lesson learned the hard way. Five years earlier, one of her predecessors had implemented the PeopleSoft HR system with more than 40 percent customization. “Every year, there are patches and upgrades,” says Fechtmann. She is “painfully de-customizing” the PeopleSoft platform (working with consultants from KPMG Consulting and outsourcer Cognizant) to upgrade to version 8.3.

After the vendor, rollout and customization choices came the hard part. Every Thursday, Fechtmann flew to Tampa to foster a close working relationship with her business partner, MetLife’s comptroller, and representatives from Met’s subsidiaries. “There would be painful meetings where we would do account mapping,” Fechtmann, a former accountant, recalls with a grimace. “New England Financial’s comptroller would say, ’We have 500 general ledger accounts, and this is how we normally post our stuff.’ MetLife’s comptroller would say, ’Well, we do it this way.’”

The IT governance board came to the consensus that decisions such as those would be made based on the greatest overall benefit, rather than simply going the MetLife way. “We weren’t going to say, ’Well, this is the way we do it, and you have to comply,’” says Fechtmann. “So we went account by account and figured out how we were going to do it. [We] really had to reengineer the business processes. [We] were really trying to define a new model of how to do business.”

The process was tedious and troublesome. “You run into both business process angst and, quite frankly, cultural angst,” Fechtmann explains. “Old habits die hard. People say, ’Well, this is the way we’ve always done it,’ and change can be difficult.” Add to that the impending loss of jobs in subsidiaries that had their own accounting departments (that work would soon be centralized within MetLife), and the tension grew. “It became much more a job of using your influencing skills to get everyone to figure out the best way to do things going forward,” she says. “The good thing was that we had a really great partnership between IT and the business. I know that can sound like a clichŽ, but I mean it sincerely. That’s how we got through it.”

MetLife and most of its subsidiaries have made the transition to the PeopleSoft EPM system and have begun to trust the numbers it generates. The next step is encouraging business partners to use it for financial analysis and business planning.

One View

Besides going public, other big changes at MetLife have influenced the move toward integration at the application level. Most important, “the chairman” wanted to shift MetLife’s focus to the customer. “With big companies like MetLife, you often don’t get treated as a single customer,” says CIO Candito, “so we needed to create an enterprisewide system that would allow us to maintain a holistic customer relationship, regardless of product sold, in order to increase customer loyalty and satisfaction.”

In the past, if a customer moved, she would have to make multiple calls to different customer service centers within MetLife — one for auto insurance, another for her dental plan and another for investments. This type of situation is clearly frustrating for customers. Having customer information stovepiped by product or business unit also makes it difficult, if not impossible, for MetLife’s sales force to cross-sell or up-sell other MetLife products.

Creating a complete customer view is a challenge facing not only MetLife, but the entire insurance industry, according to Meta Group’s Johnston. “There’s a lot of talk about the enterprise customer right now. It’s an easy thing to talk about, but it’s a hard thing to define and difficult to implement.”

The charge to create a centralized customer database came directly from Benmosche (a former IT guy himself) and his executive committee. The planning and execution rests with Candito, who says about 80 percent of his job involves application integration. He has been meeting with business customers and their CIOs to determine requirements for what MetLife has christened the Client Information File (CIF). This database will eventually house data for 100 million customers (MetLife’s target for 2010) and become one of the largest and most complex of its kind. “The complexity comes not only from the potential size of the database but also the number of different places this data will come from,” explains Johnston. “I think it will be one of the more complex information networks ever created.”

Candito and his CIO counterparts mapped out anticipated returns from the CIF for each line of business and chose Toronto-based DWL as the vendor. DWL Customer acts as middleware between the source data (from 30 different MetLife systems) and applications, creating a “gold copy” of customer information for expediency, cost and real-time availability. MetLife’s legacy systems will communicate with each other through an XML interface, and an IBM Shark storage system will store the customer data file.

Candito then developed a data model using industry standards and created a central data administration group. Prior to implementation, MetLife worked with the vendor to determine data entities and attributes, and set performance and volume metrics for benchmarking. “We got DWL to extend their model to be more supportive of what a broad-based financial service institution needs,” says Candito. “And because of the large scale of this, we knew we had to do a lot of data analysis up front.”

This spring, MetLife rolled out a pilot of the customer information file for new customers only. By June, MetLife will have implemented core components of the CIF, including the underlying software, data model and business rules for capturing client information at the front end of a client acquisition process. In November 2002, the first administrative systems in MetLife’s individual business unit will be converted to the CIF, with rollout continuing throughout 2003. Candito will begin the CIF conversion of MetLife’s institutional business systems in 2003, then prioritize and convert Met’s remaining business units.

As with the PeopleSoft conversion, application rollout will be the easy part. Business process and cultural changes have been trickier. For example, Candito and his team are considering whether to set up a centralized client management group within MetLife to address customer services currently handled at the business unit level. Business unit leaders are still antsy about the shift. “We got a small group of our key business customers together, and intellectually they agreed it was the right thing to do, but emotionally they were afraid to give up control,” Candito says. “It’s a part of change.”

Another challenge is keeping the project aligned with business goals as people come and go within MetLife over time. “Our businesses can change a lot over the period of a year, but it’s going to take more than one year to build a customer information file,” says Mark Hammersmith, senior vice president and CIO of institutional business, whose 33 million customers will eventually live in the CIF.

Architectural Integration

MetLife Senior Vice President and CTO Steve Sheinheit says integrating applications and data related to the CIF and PeopleSoft projects is the most difficult integration work under way at MetLife. Despite his attempts to downplay his role, much of the integration work begins or ends with Sheinheit. He’s responsible for the entire infrastructure — from the desktops to the mainframes — and oversees a budget of $300 million and a staff of 1,100 in seven locations. “When I came, in some ways, the hard work had been done. The buy-in was there. The understanding of the value was there,” says Sheinheit. “I just came in and did a lot of execution.”

Early on, Sheinheit established an enterprise architecture program and governance council with representatives from each of MetLife’s business units. “We had to start by bringing together representatives from the business and technology groups to figure out how we were going to establish a whole greater than the sum of its parts for this company,” he says.

A set of seven information technology principles (see “Seven Habits of the Highly Integrated Enterprise,” Page 53) to guide IT decision making was among the first things the enterprise architecture council developed. “It was the set of values on which we wanted to operate with regards to technology within the firm,” says Sheinheit. “They seemed like obvious ideals, but they weren’t always followed. So we felt it was important to communicate them, create an understanding of what each one means, and explain the rationale for and implications of each principle.”

The team also created a technology standards document that it updates frequently via the corporate intranet. It also published a road map for technology projects, detailing how MetLife would integrate across the five CIOs, the application development groups, the CTO and his infrastructure group, and ensure that necessary checkpoints and controls are established. “My sense is that not too many companies have been that successful in getting to the point that we’re at with this,” says Sheinheit. “In companies where technology is more decentralized, you’ll find even less of this kind of discipline, with different ways of doing things going on in different parts of the business.” At MetLife just a few years ago, you would have found just that.

Sheinheit’s team further reduced costs and improved efficiencies by integrating the data centers for acquisitions New England Financial and General American Financial into MetLife’s common data center on Memorial Day weekend of 2001. “It was a major effort and a milestone of our integration efforts,” says Sheinheit. “It’s resulted in a 40 percent reduction in costs and an increase in service for the business because they’re now working in a more mature, scalable environment.”

Sheinheit will continue to oversee most of these integration initiatives indefinitely. “When I look at our infrastructure today, I’d say we’re probably 80 percent integrated, but that’s a moving target. If I look at it next week, I don’t know if I’ll be at 70 percent or 90 percent,” he says. “Maybe the answer is that we’ll always be around 80 percent integrated. There will always be things we’ll continue to look at.”

Work in Progress

Like the Metropolitan Life Tower, under renovation and enshrouded in scaffolding, the company will continue its integration efforts as part of the business reorganization. Business unit CIOs are consolidating and retiring legacy systems as part of Project LESS, a three-year legacy system simplification project with a two-year payback. The goal is to consolidate like systems onto common platforms to eliminate redundant business processes and reduce ongoing maintenance costs. An Internet self-service initiative, fueled by a $200 million investment, is also under way.

There are numerous smaller integration projects within each business unit as well. Tony Colyandro, Met’s vice president and CIO of broker/dealers and investments, and his staff are combining front-, middle- and back-office functions of all four broker/dealers. They are seeking economies of scale and looking for the best solutions within each of the brokers’ existing systems.

Hammersmith has invested $120 million over several years to create self-service portals called MyBenefits.com and MetDental that will connect legacy systems for institutional clients. Richard Small, vice president and CIO of MetLife Auto and Home, has invested $40 million to integrate the personal lines of business (worth $1.1 billion in premiums) acquired from St. Paul Cos. in 1999. These pro- jects in specific business units and across the enterprise will continue for some time.

“We still have a lot of work to be done. We have a lot of legacy systems that still need to be supported that we mask with middleware,” Sheinheit says. “Today, we’re doing a lot of front-ending things, creating a veneer that makes it look as if everything’s integrated. But over time, we’re going to get all of the pieces actually integrated.”

Meta Group’s Johnston says we won’t see a truly integrated MetLife, or any other insurance company, until at least 2006 or 2007. Still, MetLife’s long-term investment in integration is exemplary, he says. “MetLife put a stake in the ground and made the decision that they’re going to have to do this or be ineffective,” he says. “That shows a great degree of moral courage on their part to invest in something where they may not necessarily see a real return on their investment in 12 or 18 months.”

For Cavanagh, the enduring nature of MetLife’s integration efforts is both logical and welcome. “Integration is an ongoing challenge because we’re always going to be adding new business, and people are going to keep coming up with new ideas in our existing business,” he says. “We’re going to find more and more things that belong together.”