Previous Next

Articles

March 2008

In 2003, John Golden, the chief information officer at CNA, the fourth largest insurance company in the U.S., had completed the consolidation of applications, had centralized more than 20 IT groups, and had trimmed systems maintenance costs. He then put a five-part strategy in place to make sure IT could put a lot of muscle in the business units, thus moving CNA forward. Golden's strategy to improve business processes includes having the right partners, doing proper program management, using more comprehensive solutions, making more use of portfolio management, and sticking to a comprehensive delivery system.

 

For the past four years, Golden has stuck to his IT strategy for CNA. "Its purpose is to make sure we do the right thing for the business. We've been evolving from an organization that did $10 million IT projects to one that does $100 million complex IT projects that will transform the company. These projects involve a business process change in addition to a change in technology. We've looked at it as an investment which would yield a return."

 

Enterpriseleadership.org recently sat down with Golden to talk about the challenges of carrying out some parts of his strategy. He oversees a centralized IT organization with 1,100 employees, and manages relationships with several partners and offshore IT resources.  In 2006, CNA had $10 billion in revenues and assets of more than $58 billion. Here's what he had to say:

 

EL: What were some of the things you looked for in finding the right partners and in building relationships with them?

 

JG: We knew that the increase in spending was a big risk. We've found two good partners -- Accenture and IBM. We wanted no more than three partners. It took us longer than we thought to find the right ones. We had to mix their style with ours, to have the right contractual terms, and then to allow them time to deliver. We quickly got a sense of each other's disciplines, and risk tolerances. With one of our partners, we now have a deep risk sharing and reward sharing. It took us a year before we could get here.

 

EL: How would you rate your program management imitative?

 

JG: We've hit a homerun with our world class program management. Our project management office is extremely strong. We have the right measures in place for consistently delivering high quality services. That's what I wanted to make sure we did. We're still maturing at the high end of a $100 million program. It's more on the execution piece in the organization with our partners. We have good methodology, good tools, and good expertise. We've really developed a good skill set there. We have good leadership here because our project management office has been very strong.

 

EL: What changes have you made to your portfolio management strategy?

 

JG: In 2004, we introduced our portfolio management strategy to make sure we had the right portfolio structure and the tools to support that. It was an area that  became problematic because of priority issues. We deserved it.  We got the right structure in place that year and the noise went away.

 

We've gotten better at the way we prioritize, the way we interact with the project team, the way that we schedule our work and the way we make decisions around the investments. We've really improved our day-to-day actions. My organization has also had a strong relationship with the executive committee, including the chairman. My peers have given me a lot of good direction. As a result, portfolio management really excels for us now.

 

EL: What changes have you made to your delivery framework so you can handle larger projects?

 

JG: We have a comprehensive delivery framework which has enabled us to be more consistent with what and how we deliver. I thought we really would've excelled in this space. I'm disappointed in my performance here.  If you measure our consistent framework on the same scale with CMMI, we do very well. Our delivery rates are in the mid to high 90's as far as scope, schedule, and budget. We've driven a lot of cost out of it to be more effective. We're still relying too much on brute force and too many heroes. However, I appreciate what the employees are doing in that space.

 

I measure our credibility based on how well we're delivering what the chairman wants. Is it giving us the ability to do things to help the company innovate? Are we able to take on large projects which would move us beyond the realm of IT? Our credibility is very good. However, I'm concerned that as I stretch my team to take on more things outside of the pure IT umbrella, my delivery framework isn't as rock sold as I 'd like it to be. I've started to stress that.

 

EL: How do you plan to solve this problem?

 

JG: The owners of the various processes need to make sure they really do have a comprehensive set of measures around those processes. We're very good at the global level. If you look at our scorecard, you'll see that our utilization, our cost of transactions, and the split between maintenance and development all have good numbers. People talk a lot about agile development. Our problem is agile development hitting the big time. We're pretty good at agile development on $5 million projects. When it comes to agile development on a $100 million project, I don't know of many organizations that are really good at it. As you begin to scale, people seem to get less disciplined and less agile. We can't do that. Our challenge is how do we be agile but disciplined on a larger project scale? Each of the process owners has to be able to work together to make sure they understand the new demands and new expectations. Can they scale it? It's not a resource issue, but a process issue.

 

EL: What increases have you made to your budget for new projects, and what new projects are on your drawing board?

 

JG: Between 2001 to 2005, our IT budget decreased each year. From 2005 to 2006, 2007, and going into 2008, our IT budget has been growing to support new initiatives. So, we've drastically been increasing what we spend on the business.  For 2006, we'll spend, on a cash basis, more than 50 percent of the budget on new work.

 

Our new projects for claims and underwriting will greatly transform our business. In underwriting, we're completing our work in small business and attacking the middle market. We've had a lot of work going on in predictive modeling and capital modeling. We have a major financial systems roadmap imitative which will take our information management system, called Merlin, to the next level.

 

EL: How does your IT organization work with the business units?

 

JG: Some people would call it a federated model, but I wouldn't give it that definition. For investments of several million dollars, I have a team of directors who deal with leaders from each line of business. The goal is to manage those investments and to help these leaders make good investment decisions. We give the leaders the allocations.  That's the day-to-day stuff. I don't monitor those decisions daily. I do expect to see a certain return from each investment.

 

EL: What is your corporate governance model for IT?

 

JG: The chairman, the CFO, the head of our property casualty business, and me make our major investment decisions, which comprise the bulk of our money. The various business units present their investment ideas to us. We decide how we want to make those investments, which aren't seen as IT projects, but as investments in CNA. They deal with process transformation, the technology piece, and any organizational components.

 

EL: Can you talk what's happening with Merlin, CNA's business information dashboard?

 

JG: Merlin, which is built on Oracle and a Business Objects,  is a scorecard of standard metrics for the business. It's clearly the information repository at CNA which helps us to determine where the business goes. We create and we align around a given metrics. The production data in Merlin helps the management at CNA to measure how our field operations are performing based their production targets.

 

The third generation of Merlin will be available in 2008.  It will have an additional complement of financial metrics, and operational metrics which can provide the cost per transaction. We've been building and maturing those metrics. Once the metrics reach a level of maturity, we put them into Merlin.

 

EL: How do you evaluate senior IT professionals to determine if they are a good fit for CNA?

 

JG: I measure people on capability, character, and chemistry. The first is usually a given. We place a lot of emphasis on ethics starting from the chairman on down. Chemistry is something we've started to focus now. We did a major turnover in the marketplace between 2000 and 2003. Now that we've proven ourselves, we can more selective about the chemistry rapport we want from our employees.

 

CNA is all about teamwork.  To evaluate my direct reports, I enlist the help of an outside psychologist to conduct behavioral interviews and to assess the fit with CNA based on those interviews. My goal is to hire people who understand the values that are important to CNA and who understand how things get done at CNA.

 

EL: Do you have a formal process for driving innovation?

 

JG: CNA is a conservative company. We don't get wrapped up in having a chief innovation officer. Of course, innovation is critical to what we're doing here. We do put a premium on it. We believe we've been innovating all along. CNA has gone through the fix, build, and lead phase. We're now building some core competencies so we can lead. Our goal is to differentiate ourselves from our competitors. Our leadership team faces the challenge of how do we bring out those good ideas to move the company forward?

 

In IT, we use Six Sigma and the IT Infrastructure Library. Unlike some companies, we don't have a dedicated team of Six Sigma Black Belts who we call upon collectively for projects.  Instead, we've distributed our certified professionals throughout the IT organization. For example, when we began our claims transformation projects, we asked one of our Six Sigma Black Belts to lead the effort.

 

EL: What do you get out of your involvement with the Executive Club of Chicago's Technology Group?

 

JG: I've gotten many things from being involved with this organization. It allows my staff to learn new things by attending the 90-minute talk sessions. I get to help set the agenda for these sessions, as well as other technology things we do. CNA also gets a lot of recognition in the market place. It's also a great recruiting tool for me. It helps me to network with other Chicago CIOs, who want to promote the marketing of IT in Chicago.

--

Additional Reading - Sponsor Links:
Understanding ITIL® Service Portfolio Management and the Service Catalog
Managing the Business of IT: Maximizing the Power of Service Resource Planning, the Next Step in Business Service Management

 

Elizabeth M. Ferrarini is a writer from Boston, Massachusetts. Reach her at elizabethferrarini@yahoo.com.

| More
136 Views 0 Comments 0 References Permalink Tags: article, best_practices, it_management

Many CIOs have faced the challenge of how to align their technology investments with overall corporate strategy. eBay, the world's largest e-commerce auction site, faced a different type of challenge. From day one,  eBay had to create its own world to be successful. Technology plays an integral part in shaping the direction and the growth of eBay's business. One could favorably argue that eBay is about technology. Because of its business model, eBay had to build its architecture from scratch and to devise new best practices to support it. James Barrese, eBay's vice president systems and architecture, says that building many key components can create new problems that must be addressed without the legacy safety net to fall back on. enterpriseleadership.org recently sat down with Barrese to talk about how his group maintains eBay's agile architecture and scalability to support a 223 million user base, which grows by 130,000 new users each day. Here's what he had to say:

 

EL: Can you describe what your organization  does?

 

JB: The architecture team works on the OS for eBay. You can imagine that we have many applications that serve different purposes for the communities. We also have a large and sophisticated operational environment.  The systems team builds the eBay OS, which comprises security, kernal layers, management, infrastructure, and logging.  The entire reusable software infrastructure makes it easier and faster to develop applications.

EL: How do you align with IT?

 

JB: We're part of eBay's product development organization. eBay differs from most e-commerce organizations because the software we create becomes the business. We align in multiple respects with the strategic roadmaps. We, however, run a separate technology strategy, which aligns with both the business and the needs of the product development organization, as well as the needs of the operations teams. Every year we bubble up a few major initiatives that become the strategic enhancements we need.

 

EL: Do you have a formal process for  driving sustainable innovation?

 

JB: Absolutely! We've done that across several different dimensions. On one hand, we have very far-reaching breakthrough types of innovation. These things come out of our eBay research lab. We have a separate bucket of technology investments, which we call our headroom process. It's a combination of making investments in new technologies, but it's also about going through and improving our infrastructure where we know we're going to need scaling. We've done that process to drive up the eBay infrastructure. Every year we have a certain amount for our software development budget. We carve out funds for infrastructure investment, which we then use to innovate on our platform.

 

EL: Can you describe your architecture platform, which provides SOA components, security management, as well as other technologies?

 

JB: Those are a few big areas. We've created a very nice, zone infrastructure. It's a full application and architecture for making it easy to create services within services. It also makes it possible to plug things in, so we aren't wedded to one particular appliance or protocol. That's  going to be a major enabler of the business going forward.

 

We have  many dimensions to security management, everything from writing the professional rosters to locking the front doors. eBay, being one of the largest Web sites in the world, gets targeted constantly. Over the years, we've hardened our system to be able to handle everything that has come at us. As for analytics, we've one of the world's leading data marts and analytics infrastructure. It enables us to process data, to have insight into actions, and to be able to drive business strategies from the data we're saving

 

EL: Any other technologies that are part of the  architecture?

 

JB: We use a lot of Solaris. We built our own search subsystem. We use Java very heavily. We built a whole Java framework on top of hung Java. We work with IBM. We have some Cisco in the shop. We have a lot of enterprise tools. Where there are limits to those tools, we've built up additional layers of infrastructure that really solve many of the problems we have at our scale. Two of our other vendors include Oracle and Hitachi.

EL: Can you describe how your group works with  third-party developers?

 

JB: That program is in my systems group. With the new Web generation, we want to strike the right balance between being open and encouraging entrepreneurship, while having oversight to ensure quality. For example, we have an online directory of different applications people have built. We allow the open community to rate those applications. We then raise visibility or lower visibility based on feedback. It's an inherent built-in mechanism for those third-party developers to create quality.  It's very much like any community-driven process where we try to get away from business-oriented regulation and, instead, to embrace the community to both reward great applications. However, we also make sure we limit our quality products. We find it's a better and a more scalable mechanism for an environment like ours, which is in constant flux. It's a better oversight method than any other rigid process that we might have invented.

EL: Do you use  any quality practices such as Six Sigma or COBIT?

 

JB: We've integrated the best parts of those two methodologies and created an eBay product development lifecycle (PDLC). We have a very structured process where we have requirements, scopes, projects, and everything for QA. Because we built this eBay OS, we have invested heavily in automating our entire development and rollout process. We can easily manage 100s and 100s of developers, and 100s and 100s of projects going through the same PDLC. We can allow a small one- or two-person development team to go through the same process and tools as a 50- person development team. It scales. We've tried to take a lot of the redundant human element out of it. As a result, we really enforce those best practices and those standards around development and quality, and management of our release process, as well.

 

EL: How often do you release your  OS?

 

JB: We roll out our entire site every two weeks. One week we roll out our North American site, and the following week, we roll out all the international sites.  We constantly re-roll our entire stack. Every single one of those releases has features, additional functions, and new applications going out.

EL: Because you're mostly a build versus a by shop, what advice what you would give to IT shops steeped in integrating custom products with third-party solutions?

 

JB: I'd tell them to be careful not to be a fast food order taker and just to build what you hear the business asking for. When you're building custom software, you're making a big decision and a very significant investment in this process. It has to be long lived. Don't deliver what you hear. You need to be a partner and to craft solutions that meet the heart of what the business is after. On the other hand, you have to balance the costs of custom development to make sure you're hitting that core strategic need of the business. You also have to craft that solution with them and strike the right balance.

EL: While eBay continues to scale,  how do you remain agile?

 

JB: We've been investing in a program called Nexus. It's an infrastructure that has a lightweight, satellite component. This eBox allows a very small team to be very rapid and nimble. Likewise, it's the same software stack and infrastructure that we can build an application that is serving 100 million users. This vehicle enables a small innovation team to crank out something, such as a marketing application, very quickly. If you do a 100 of these, then a few of them are going to be wildly successful. You're going to need to scale them. Rather than rewriting them, we're putting them on an infrastructure, or the eBay platform, which is going to allow us to scale to an enterprise class, 24 by 7 highly available system.  We've completed a few releases of this. A team of PHP developers Germany came in fresh to our technologies. They had rapidly learned ebox, We developed an application faster than we could previously. It was very successful. That's how we've extended our infrastructure for rapid innovation and very mature high quality applications as well.

EL: What kind of a development  program do you have for your staff to continue to innovate?

 

JB: We do a few things. A lot of our projects involve trying to solve problems related to our scale. It's a job of making it better.  We have people who find  fixes to low-level operating system issues and JBM's. They're contributing their fixes back to the open source community, as well as pushing vendors to improve their solutions.

 

I can't tell you how many times our junior-level engineers find solutions to a major vendor's product. On the job, you have to be an innovator. You have to continue learning and extending your skills. eBay also has a development program for tuition reimbursement for education. We also run internal training programs.

 

We run all kinds of innovation programs. We call them Innovation Days and Skunkworks programs. I'm on the scoring committee for the next one we're going to do. We allow people to create any kind of idea or wildly creative solution they might've thought of. Typically, they build it on the eBay APIs. We reward and promote those different applications. One of these applications went live on the site. For example, a product, called eBay Countdown, came out of Skunkworks.  It's a Web 2.0 Flash-based application that let's our community watch items as the auction counts down. That's an example that went from concept to live on the site quickly. eBay To Go, which is another one of our most popular widgets, is another example of an innovation.

EL:  What are you doing to help large sellers become  more efficient?

 

JB: We focus constantly on seller efficiency. If you think about the entire developer's program, we have more than 55 percent of our eBay listings coming through the APIs. We've built an entire set of software, infrastructure, and APIs to allow sellers to automate their businesses. You can see just by the volume over the years. The sellers are more efficient.

 

EL: How transparent is your technology to your  constituents?

 

JB: We're very transparent in what we do. We're clear on the levels of the investments we made in business projects, in our infrastructure, and our research and development lab.  Even within those infrastructure investments, we have transparency about what we're doing. It has to be aligned with what the business wants.  The Nexus program drives both innovation and flexibility. It as much a technology need as it is a business need.  How do we innovate faster? How do we have shorter time to market? How do we lower costs? Any business person is going to be excited about any solution that can do these things.

 

EL: Have you gone to server  virtualization or grid computing?

 

JB: We are doing a lot in virtualization with a few different virtualization technologies.  Some parts of our system are running in a virtualized environment.  Our quality assurance has seen dramatic benefits from virtualization. Not all of our product environments will benefit from virtualization.

 

We've invested heavily in grid computing. For example, we've automated our entire release process. When a software developer checks in his or her code, an entire workflow gets merged to the main system, and then goes through regression testing on QA. We have an automated environment that rolls it out and distributes it to our site. On top of that, we have all types of analytics and problem detection. We can see from release to release if there is a dramatic variation in how long a process is taking. That will flag problems for us. A number of automated tools help us to triage and to influence what the problem is.

 

Author: Elizabeth M. Ferrarini -- She is a technology writer  from Boston, Massachusetts. Reach her at elizabethferrarini@yahoo.com.

| More
114 Views 0 Comments 0 References Permalink Tags: article, best_practices, it_infrastructure