Liberty Mutual Insurance has turned out to be the company to watch in the fiercely competitive global insurance industry. Since 1996, this sixth largest property and casualty insurer in the U.S. has seen its annual revenues triple to about $26 billion. Today, Liberty Mutual ranks as the sixth largest property and casualty insurer in the U.S, has $85 billion in assets, and has more than 40,000 working across 900 officers worldwide.
Because Liberty Mutual's offers a variety of lines of business, ranging from personal auto, health and life insurance to commercial property insurance and workmen's compensation, the company has adopted a highly decentralized business unit structure. Although this is a powerful business model, it also has presented many challenges to Stuart McGuigan, Liberty Mutual's global CIO. Enterpriseleadership.org recently sat down with McGuigan, who was the former CIO of the $32 billion Medco Health Solutions, to discuss how IT delivers the technology required by the business units to meet marketplace demands. Here's what he had to say.
EL. Can you describe the structure of your IT organization?
SM. We have 3,000 people in a centralized IT organization. We don't bother to draw a dotted line to the businesses. We have established business CIO who report to me, but who sit on the staff of the president of the business units. These CIOs attend staff meetings, strategy sessions, and offsite meetings. Their entire focus and their passion include how to use technology to make the business more successful. This structure keeps everyone talking to each other. Although these CIOs work for me, I don't want the CIO supporting the personal markets business concerned with what the CIO of the commercial markets is doing, unless an initiative can help a CIO do something better, faster, and less expensive for a business unit.
EL. What factors can you attribute to Liberty's rapid financial growth to?
SM. We can attribute our success to Ted Kelly, our CEO, who came to us in the early 1990s. He discovered the secret of growing a large company with significant diseconomy of scale.
When I came to interview at Liberty Mutual with Kelly, I began to understand why management capacity has driven Liberty's growth rate in the insurance industry. Kelly empowers general managers to exploit everything from niches to significant markets with a tremendous amount of operational latitude, and with a certain amount of standards to minimize financial risks.
To this end, the executive leadership of each Liberty Mutual business unit determines how that entity goes to market, how it is organized, how it approaches its marketplace, and what type of technology it uses. As a result, Liberty has exploited numerous niches in midsize businesses by acquiring companies, both in the U.S. and outside the U.S. These companies very quickly adopted Liberty's business model. In every case, we've succeeded in the business case. This's something many large companies struggle with.
EL. What does the decentralization of management authority mean for IT?
SM. If you don't do something differently, then every group and every application team decides its own technology. You end up heterogeneous technology environment, which leads to cost. You can't sacrifice the responsiveness and effectiveness. Over time, we've been making Liberty more than the sum of its parts. We looked at technology in a different way starting layer by layer and said: Which technologies provide no value, which technologies offer added benefits, or which technologies are different across the various business units. We gradually moved up in technology stacks. Does it really matter what Unix server people use? Some people initially said that it did matter. We determined that some applications could you use certain types of Unix servers. These three servers became a standard. We said if you want to use these standard servers, then you can't take 100 days to complete the installation, you have to do it in 10 days and at this cost. If the business units could derive enough business value from using unique technology or custom technology, we told them to for it.
EL. Can you go into more detail about how you built flexibility into the decision to go with standard technology?
SM. Some IT organizations tell the business units that they can only use standard technology, or what IT has stamped as standard. So how can they benefit collectively when no one benefits individually? That's the dilemma of dealing with some types of enterprise architecture strategies. We've avoided that by saying we're going to standardize up to the point of indifference and beyond they can choose. We're not continually going to challenge their decision.
We're trying to standardize where it doesn't make a business difference and to allowthe businesses to decide what technology will enable them to be successful. Because they bear the entire cost, which is the biggest part of their business resource decision making, this IT philosophy fits in with the way they manage the business. It's not something external.
EL. Do you see any difference between the way business units make IT investment decisions versus any other operational business investments?
SM. Usually IT people believe there are two separate entities -- IT and the business. IT is a business function that manages technology. The thinking that does into the decision to hire 20 underwriters shouldn't be any different from the thinking that goes into the decision to spend a million dollars on an underwriting system. Both decisions will produce different results, but you're looking at the same variables. Getting that balance right is the secret to realizing some of the benefits of being a large-scale enterprise, but you don't want to sacrifice your ability to exploit smaller business niches, such as startups.
EL. What aspects of IT do you need to improve and why?
SM. First of all, look at our decentralized business structure. Each business unit organizes itself around business, products, and customers. For example, the real driver behind the retail business isn't the product, but the retail marketing operation. Think about the technology that a sale-driven and a direct marketing retail organization needs versus the technology that a business-to-business organization needs to support customers such as IBM! Think of the differences in business processes, product requirements, and customization.
If they were truly separate companies, no one would argue they should have the same architecture. Several standalone, billion dollar businesses comprise Liberty. Although we are a large company, we found that the cost of IT isn't our issue. In fact, our cost of IT is really a percent of premium or revenue, which is on the low side.
Our issue is this: like every other insurance company, we need to improve out time to market. If we save 10 percent of our IT cost, but we don't improve our business, then we'll be out of business. If a good IT project is a 2 to 1 or 3 to 1 investment over time, then IT spend too much time trying to make that $1.95 or not enough time making sure we get the $2.
EL. What's wrong with the discussion about aligning IT with the needs of the business?
SM. The conversation about business and IT alignment infuriates me. Of course, I try to be patient if someone wants to talk about it. If you step back and think about it, what does it mean to have IT projects that aren't aligned in the business? You're spending money on things that don't produce a business result, not even in theory. Don't do it! What other business area makes capital investments and refreshes its equipment with no idea of how it's going to contribute to any business results? Only IT does that. We have to get out of that cycle. Pharmaceutical companies, for example, have a replacement cycle for when their tablet-making machines will produce lower quality products or breakdown. There is a cost associated with replacing those machines. It comes down to an optimal point to which you maintain and replace those machines. It's not vendor recommended because it's end of life.
Three years a go, we had a proposal to replace all of our Intel servers with smaller departmental servers. The vendor said they were at the end of their life. I questioned what made them end of life -- a decrease in performance, many outages, or high energy costs. I said we shouldn't do it. My decision shocked everyone because we had the money, and because we did it every three years. I told my staff to keep track of the mean time between failures, keep track of performance, and keep track of newer versions of the software. If we started to see any of these things go south, then we'd have the beginning of a business proposal to spend the capital to make that replacement.
Three years later, we're now replacing those servers because of those changes. We had the highest level of availability, the lowest cost of any one else in recent benchmarks. We accomplish that because we didn't fool with the systems Changing your environment can produce defects in reliability. We put the maintenance on a different footing, made different decisions, and made sure our staff understood that we weren't walking away from technology. Instead, we were investing technology dollars in an area that could produce the biggest return and that could avoid significant issues. We didn't fall prey to follow the leader by replacing technology and making IT investments just because it's common practice.
EL. Can you describe your governance model?
SM. We have a project management office for governance. We need to follow guidelines for both Sarbanes Oxley and for good practice measurement. We need the ability to see whether we're improving our processes, improving our quality, or improving our productivity. We have mandated criteria from product development, such as phase gates or stage gates, project initiation, design readiness, development readiness, testing readiness, production readiness, and the new post-implementation evaluation. The milestones we assign to a project require us to go through a structured set of questions before we can move to the next level. These questions address how prepared we are to move to the next level. We don't wind up 90 percent into a project only to realize we don't have a chance of success. We catch issues and problems early as part of the evaluation. That's how we exercise quality governance.
EL. How do you look at projects that will benefit the entire organization?
SM. We have a set of projects we do across the enterprise. For example, we have some infrastructure investments that will help everyone. We'll discuss these investments with the heads of the business units. Our chief executive officer who reports to the CEO drives many capital IT investment decisions. For example, last year, he spearheaded a very aggressive project to enable voice over IP to our 300 offices. It was a technology needed by out retail business, but it is clear to us that we could use that flexibility in all of our operations in the near future. Rather than do it piecemeal, we did it as an enterprise project.
EL. How do you justify IT projects for each of the different business units?
SM. We justify projects on a business-by-business basis. For example, two of our business units have invested in customer centric data management, which will them provide a 360-degree view of the customer. This's a new capability in insurance. It is also good for marketing and good for relationship management. These business units need that technology as fast as they can jointly build it together.
Meanwhile, the commercial business that includes worker's compensation and general liability for large companies, is a mature business. We've put much good technology in place for this business unit. This business unit's projects are all about reducing the cost of business process and IT to provide more pricing flexibility.
The small commercial and personal line business unit goes through independent agents as oppose to selling directly to end customers. These independent agents handle relationships with multiple insurance companies. They have a different business model than other business units. They also use an entirely different set of tools. They also have a whole set of different issues. They're in very early stages and they're trying to build some basic integrated capabilities and consolidating some operations. They have a different set of needs.
Author: Elizabeth M. Ferrarini - She is a technology writer from Boston, Massachusetts. Reach her at elizabethferrarini@yahoo.com.
