CASE – 1 The 2004 Athens Olympics Network: Faster, Stronger—and Redundant
Claude Philipps, program director of major events at Atos Origin, the lead IT contractor for the Olympic Games, likes to be prepared. “We were ready before August, but we were still testing, because we wanted to be sure that every stupid thing that can happen was planned for,” Philipps said. “In a normal IT project, we could have delivered the application to the customer almost eight months earlier.”
But the Olympic Games was far from a normal IT project. The deadline was nonnegotiable, and there were no second chances: Everything must work, from the opening ceremony on August 13 right to the end, said Philipps, whose previous experience includes developing the control system for the world’s first computerized nuclear power plant.
With all that pressure, Philipps’s team was doing its utmost to ensure that the network would not fail. They were building multiple layers of security and redundancy, using reliable technology, and then testing it rigorously.
In the weeks before the games, the team went through two technical rehearsals in which 30 Atos Origin staffers put the network through its paces. The team spent a full week stimulating the busiest days of the games, Philipps said, dealing with “crazy scenarios of what might happen in every area: a network problem, staff stopped in a traffic jam, a security attack…everything that might happen.”
The rehearsals were intended to test people and procedures as much as the hardware and software. That was important because the IT organization Philipps built for the Athens Olympics grew from nothing to a staff of 3,400 in less than three years.
The two major components of the software that were run over the Olympic network were Atos Origin’s GMS (Games Management System), a customized suite of applications that acts as kind of ERP for the Olympics, and the IDS (Information Diffusion System).
GMS ran on Windows 2000 servers in Athens, an upgrade from the Windows NT 4 used at the Salt Lake City games in 2002. “We’re not using sexy technology,” Philipps said. “The main goal for us was to reduce the amount of risk.”
Together, GMS and IDS imposed exacting requirements on the network. GMS was, among other things, used to manage access accreditations for the games, so security was vital. Speed, too, was important: Philipps’s goal was to have the results on commentators’ screen 0.3 seconds after the athletes had crossed the line, complete with rankings, statistics, and biographies—everything that helps commentators during a live broadcast.
Yan Noblot, information security manager at Atos Origin, said the key to that was to build in redundancy—and lots of it. “We doubled everything, because we needed 100 percent availability at games time,” he said.
And when he said everything, he meant it. There was backup redundancy for the routers and switches at each site, the datacenters that processed the results, and event the PCs on the desks in the control room.
Too keep things orderly, Atos designed three different LAN configurations: one for the largest venues, including the Olympic stadium and the water sports center; another for mid-size venues such as the equestrian center; and one for the many smaller venues.
Atos used VLANs both to simplify troubleshooting and to limit damage if anyone managed to break into the network. There were separate VLANs for the commentator information system, information diffusion applications, and the game management system. Technical services, directories, management and monitoring, and the on-venue results system each had their own VLANs too, sometimes several per venue for the same function.
“The purpose was to segment the traffic so we could monitor it and contain potential issues,” Noblot said. “If someone brought in a virus, it would be contained on systems on the same VLAN and could not spread to other VLANs.”
Event results and data from the games management system were stored in two physically distant data centers hosted by OTE, which also supplied the SDH network. The primary data center was located near OTE’s headquarters in Marousi, just across the main highway from the Olympic stadium; the other was another several hundred miles away, still in Greece but in a different earthquake zone.
What makes the Olympic Games a unique project is that the athletes aren’t going to stop running just because the server does. As Philipps said, “When we speak about fixing something, it might be a work-around or a decrease of functionality, but the key thing is that the show must go on.”
1. Could the 2004 Athens Olympics have been a success without all of the networks and backup technologies?
2. The 2004 Olympics is a global business. Can a business today succeed without information technology? Why or why not?
3. Claude Philipps said dealing with “crazy scenarios of what might happen in every area: a network problem, staff stopped in traffic jam, security attack…everything that might happen,” was the reason for so much testing. Can you think of other business that would require “crazy scenario” testing? Explain.
CASE – 2 Argosy Gaming Co.: Challenges in Building a Data Warehouse
When you’ve got half a dozen riverboat gambling operations, it’s important that everyone plays by the same rules. Argosy Gaming Co. (www.argosycasinos.com), with headquarters in Alton, Illinois, and a fleet of six Mississippi riverboat casinos, had decided that bringing all customer data together would enhance management’s view of operations and potentially help strengthen customer relationships. To accomplish those goals, though, the company needed to access a variety of databases and develop an extract, transform, and load (ETL) system to help construct and maintain a central data warehouse.
Jason Fortenberry, a data-warehousing analyst, came aboard at Argosy just as the company’s data warehouse project started in 2001. His job was made easier, he says, by the adoption of Hummingbird Ltd.’s Genio ETL software tool, which helped bridge systems and automate processes. But like others going through such projects, he learned the hard way that preparing for the ETL process is just as important as having the right software.
The riverboats each had unique and incompatible ways of defining a host of operational activities and customer characteristics—in essence, the floating casinos were each playing the same game but with different rules. But those problems remained hidden until reports from the company’s data warehouse began to turn up inconsistent or troubling data. That’s when Fortenberry and his staff discovered conflicting definitions for a wide range of data types—problems he wished he had identified much earlier. Fortenberry’s troubles—and his successes—are typical of ETL, the complex and often expensive prelude to data warehouse success.
ETL is often problematic because of its inherent complexity and underlying business challenges, such as making sure you plan adequately and have quality data to process. Analysts, users, and even vendors say all bets are off if you don’t have a clear understanding of your data resources and what you want to achieve with them. Then there are choices, like whether to go for a centralized architecture—the simplest and most common configuration – or a distributed system, with ETL processing spread across various software tools, system utilities, and target databases, which is sometimes a necessity in larger, more complicated data warehouses. Even if you navigate those waters successfully, you still need to ensure that the ETL foundation you build for your data warehouse can meet growing data streams and future information demands.
As the term implies, ETL involves extracting data from various sources, transforming it (usually the trickiest part), and loading it into the data warehouse. A transformation could be as simple as reordering the fields of a record from a source system. But as Philip Russom, a Giga Information Group analyst, explains, a data warehouse often contains data values and data structures that never existed in a source system. Since many analytical questions a business user would ask of a data warehouse can be answered only with calculated values (like averages, rankings or metrics), the ETL tool must calculate these from various data sources and load them into the warehouse. Similarly, notes Russom, a data warehouse typically contains “time-series” data. The average operational application keeps track of current state of a value such as a bank account balance. It’s the job of the ETL tool to regularly add new states of a value to the series.
For his year-long ETL project, Argosy’s Fortenberry says Hummingbird’s Genio Suite, a data integration and ETL tool, quickly became the project’s “central nervous system,” coordinating the process for extracting source data and loading the warehouse.
But for Argosy, getting all that data into the warehouse didn’t produce immediate usable and dependable results. “ The lesson was that people thought that they were talking about the same thing, but they actually were not,” says Fortenberry. For example, riverboats calculated visits differently. One riverboat casinos would credit a customer with a visit only if he actually played at a slot machine or table. Another had an expanded definition and credited customers with visits when they redeemed coupons, even if they didn’t play. So identical customer activity might have on riverboat reporting 4 player visits and another reporting 10. “This type of discovery was repeated for everything from defining what a ‘player’ is to calculating a player’s profitability,” says Fortenberry.
IT played a lead role in identifying problems and helping to hammer out a consensus among the business units about how to define and use many categories of data, he says. Now, the data warehouse is running smoothly and producing dependable results for business analysis and management reporting, so the number of problem-resolution meetings has dropped dramatically. Still, Fortenberry reckons that three-quarters of the meetings he attends nowadays have a business focus. “For our part, we now know better what questions to ask business users as we continue with the data warehouse development process,” he says.
1. What is the business value of data warehouse? Use Argosy Gaming as an example.
2. Why did Argosy use a ETL software tool? What benefits and problems arose? How were they solved?
3. What are some of the major responsibilities that business professionals and managers have in data warehouse development? Use Argosy Gaming as an example.
CASE – 3 Allstate Insurance, Aviva Canada, and Others: Centralized Business Intelligence at Work
The most common approach to business intelligence is to assemble a team of developers to build a data warehouse or data mart for a specific project, buy a reporting tool to use with it, and disassemble the team upon the project’s completion. However, some companies are taking a more strategic approach: standardizing on fewer business-intelligence tools and making them available throughout their organizations even before projects are planned. To execute these strategies, companies are creating dedicated groups, sometimes called competency centers or centers of excellence, to manage business-intelligence projects and provide technical and analytical expertise to other employees. Competency centers are usually staffed with people who have a variety of technical, business, and data-analysis expertise, and the centers become a repository of business-intelligence-related skills, best practices, and application standards.
About 10 percent of the 2,000 largest companies in the world have some form of business-intelligence competency center, Gartner Inc. and Howard Dresner says. Yet approaches vary. While most are centralized in one location, a few are virtual, with staff scattered throughout a company. Some are part of the IT department – or closely tied to it—while others are more independent, serving as a bridge between IT and business-unit managers and employees.
Allstate Insurance Co.’s Enterprise Business Intelligence Tools Team is responsible for setting business-intelligence technology strategy for the company’s 40,000 employees and 12,000 independent agents, says Jim Young, the team’s senior manager.
Based in Allstate’s Northbrook, Illinois, headquarters, the center was created earlier this year by consolidating three groups built around separate business-intelligence products used in different parts of the company. The center serves as a central repository for business-intelligence expertise, providing services and training for Allstate employees, and is developing a set of standard best practices for building and using data warehouses and business-intelligence applications.
“That way, we can execute on a common strategy,” Young says. The center maintains a common business-intelligence infrastructure and manages software vendors and service providers.
At Aviva Canada Inc., a property and casualty insurance company, the primary role of its Information Management Services department is to bridge the communication gap between business-intelligence-tool users and Aviva’s IT department.
“Business intelligence isn’t a technology issue. BI is a business issue,” says Gerry Lee, information management services VP. Centralization is critical, because Aviva’s goal is to grow by 50 percent over the next five years, partly through additional acquisitions, Lee says. The center also impacts the company’s numerous customer relationship-management initiatives. “We couldn’t get into CRM until we had solid data-management and business-intelligence capabilities,” he says.
Cost reduction is often the driving factor for companies to create competency centers and consolidate business-intelligence systems. Standard technology and implementation practices can reduce the cost of some business-intelligence projects up to 95 percent, says Chris Amos, reporting solutions manager at British Telecom. BT established a center of excellence around Actuate’s reporting software three years ago and is developing business-intelligence systems for the telecommunications company’s wholesale, retail, and global services operations.
Despite the potential savings, funding can be an issue for creating and running business-intelligence centers of excellence. Start-up costs for a business-intelligence competency center can be $1 million to $2 million, depending on a company’s size, Gartner’s Dresner says.
Many believe the payoffs are worth it. General Electric Co.’s energy products business formed its Business Data-ModelingCenter of Excellence last year to improve data-management and business-intelligence practices for GE Energy’s 8,000 employees. That has helped the business move beyond simple reporting o financial and supplier data to more advanced forecasting and predictive analysis.
“The data’s become more actionable. The visibility of this data to the business has brought millions in savings,” says Rich Richardson, manager of business data modeling and delivery, who manages the center. That’s a business-intelligence competency center that’s more than paid for itself.
1. What is business intelligence? Why are business-intelligence systems such a popular business application of IT?
2. What is the business value of the various BI applications discussed in the case?
3. Is a business-intelligence system an MIS or a DSS?
CASE – 4 Blue Cross, AT & T Wireless, and CitiStreet: Development Challenges of Self-Service Web Systems
When Web-based self-service is good, it’s really good. Customer satisfaction soars and call center costs plummet as customers answer their own questions, enter their own credit card numbers, and change their own passwords without expensive live help.
But when Web-based self-service is bad , it’s really bad. Frustrated customers click to a competitor’s site or dial up your call center—meaning you’ve paid for both a self-service website and for a call center, and the customer is still unhappy. A poorly designed Web interface that greets self-service users with a confusing sequence of options or asks them questions they can’t answer is a sure way to force them to call a help center.
Blue Cross-Blue Shield. For Blue Cross-Blue Shield of Minnesota (www.bluecrossmn.com), developing Web self-service capabilities for employee health insurance plans meant the difference between winning and losing several major clients, including retailer Target, Northwest Airlines, and General Mills. “Without it, they would not do business with us,” explains John Ounjian, CIO and senior vice president of information systems and corporate adjudication services at the $5 billion insurance provider. So when Ounjian explained to executives that the customer relationship management (CRM) project that would enable Web-based self-service by client employees would cost $15 million for the first two phases, they didn’t blink.
Blue Cross-Blue Shield also learned the importance of communicating with business units during the design phase of its Web self-service system. Ounjian and his technical team designed screen displays that featured drop-down boxes that they thought were logical, but a focus group of end users that examined a prototype system found the feature cumbersome and the wording hard to understand. “We had to adjust our logic,” he says, of the subsequent redesign.
AT & T Wireless. When AT & T Wireless services (www.attws.com) began rolling out its new high-bandwidth wireless networks, its self-service website required customers to say whether their phones used the older Time Division Multiple Access (TDMA) network or the newer, third generation network. Most people didn’t know which network they used, only which calling plan they had signed up for, says Scott Cantrell, e-business IT program manager at AT& T Wireless. So AT & T had to redesign the site so the customer just enters his user ID and password, “and the application follows built-in rules to automatically send you to the right website,” Cantrell says.
According to Gartner Inc., more than a third of all customers or users who initiate queries over the Web eventually get frustrated and end up calling help center to get their questions answered.
Whether a self-service application is aimed at external customers or internal users such as employees, two keys to success remain the same: setting aside money and time for maintaining the site, and designing flexibility into application interfaces and business rules so the site can be changed as needed.
CitiStreet. CitiStreet (www.citistreetonline.com) is a global benefit services provider managing over $170 billion in savings and pension funds and is owned by Citigroup and State Street Corp. CitiStreet is using the JRules software development tool to make rules changes in its benefits plan administration systems, many of them featuring Web-based employee self-service. JRules manages thousands of business rules related to client policies, government regulations, and customer preferences. Previously, business analysts developed the required business rules for each business process, and IT developers did the coding. But now analysts use JRules to create and change rules, without the help from developers, says Andy Marsh, CitiStreet’s CIO. “We’ve effectively eliminated the detail design function and 80 percent of the development function,” says Marsh. IT is involved in managing the systems and platforms, but it’s less involved in rules management, he says.
The software helps speed the development process for new business systems or features, says Marsh. For example, it used to take CitiStreet six months to set up benefit plans for clients; it now takes three months. CitiStreet can also react more quickly to market changes and new government regulations. It has used the rules development software to quickly revise business rules to accommodate the changes in pension programs required by new legislation. And Marsh says that when a client company recently added a savings plan to its benefits program, CitiStreet was able to easily develop and implement changes with JRules.
1. Why do more than a third of all Web self-service customers get frustrated and end up calling a help center? Use the experiences of Blue Cross-Blue Shield and AT & T Wireless to help you answer.
2. What are some solutions to the problems users may have with Web self-service? Use the experiences of the companies in this case to propose several solutions.
3. Visit the websites of Blue Cross-Blue Shield and AT & T Wireless. Investigate the details of obtaining and individual health plan or a new cell phone plan. What is your appraisal of the self-service features of these websites? Explain your evaluations.
CASE – 5 Avon Product and Guardian Life Insurance: Successful Management of IT Project
It’s déjà vu again at many companies when it comes to track record in using IT to help achieve business goals. Consider the following:
- At companies that aren’t among the top 25 percent of IT users, three out of 10 IT projects fail on average.
- Less than 40 percent of IT managers say their staffs can react rapidly to changes in business goals or market conditions.
- Less than half of all companies bother to validate an IT project’s business value after it has been completed.
Those are just a few of the findings from a survey of IT managers at about 2,000 companies, including more than 80 percent of the Fortune 1,000, released in June 2003 by the Hackett Group in Atlanta. However, top-tier IT leader didn’t reach the top of their professions by being softies.
Indeed, a vast majority of them regularly rely upon hard-dollar metrics to consistently demonstrate to top brass the business value IT investments are expected to yield. That’s what sets them apart from so many of their colleagues. “Good business-case methodology leads to good project management, but it’s amazing how many companies fall short here,” says Stephen J. Andriole, a professor of business technology at Villanova University and consultant at Cutter Consortium. The lack of good project management at such companies may also lead to business units taking on IT development projects without the knowledge or oversight of a company’s IT department. Business units may initiate such “rogue projects” because they see the IT department as too slow, or a source of too much red tape and extra costs.
Avon Products. “We apply all of the analytical rigor and financial ROI tools against each or our IT projects as well as other business projects,” says Harriet Edelman, senior vice president and CIO at Avon Products Inc. (www.avon.com) in New York. Those tools include payback, NPV, and IRR calculations, as well as risk analyses on every investment, she says.
The $6 billion cosmetics giant also monitors each IT project to gauge its efficiency and effectiveness during the course of development and applies a red/yellow/green coding system to reflect the current health of a project, says Edelman. A monthly report about the status of projects that are valued at more than $250,000 and deal with important strategic content is presented to senior line managers, the CEO, and the chief operating officer. In addition, Avon uses an investment-tracking database for every IT project to monitor project costs on a rolling basis. The approach makes its easier for the company’s IT and business managers to quickly determine whether a project should be accelerated, delayed, or canceled and assists the finance organization in forecasting requirements.
Guardian Life Insurance. Dennis S. Callahan says he has “put a strong emphasis on governance” since becoming CIO at The Guardian Life Insurance Company (www.glic.com) two years ago, Callahan has done so, in part, by applying NPV and IRR calculation to all IT projects with a five-year cash flow. “The potential fallout from inaction could result in loss of market share,” says Callahan, who was promoted to executive vice president recently. So Guardian’s approach to IT investments “is very hare-dollar- and metrics oriented, with a bias toward action,” he says. Still, Callahan and his team do have a process for incorporating “soft” costs and benefits into their calculations. They do that, Callahan says, by encouraging their business peers “to discuss how an investment can impact market share and estimate how those numbers are going to change. Same thing with cost avoidance – if we invest in a project that’s expected to help us avoid hiring 10 operations staffer to handle growing business transaction volumes.”
Callahan also keeps close tabs on capital spending throughout the course of a project. New York-based Guardian has a project management office that continually monitors the scope time, and cost of each project valued at more than $100,000, according to Callahan. Guardian also has monthly reviews of variances of scope, time, and costs on all projects costing more than $100,000.
Using return-on-investment calculations to cost-justify and demonstrate the value of IT investments to senior management is only of the techniques top IT leaders use to win project approvals, says Callahan and others. “We approach everything that we do in terms of payback.” President and CEO Dennis Manning and other board members “really relate to that kind of justification,” Callahan says. “So we turn that into hard-dollar returns and benefits for application development and infrastructure investment.” “One of the biggest things we do in demonstrating value to the CEO and the board is showing that everything we do reflects the company’s business strategy,” says Rick Omartian, chief financial officer for Guardian’s IT department.
1. What are several possible solutions to the failures in IT project management at many companies described at the start of this case? Defend your proposals.
2. What are several key ways that Avon and Guardian assure that their IT projects are completed successfully and support the goals of the business?
3. If you were the manager of a business unit at Avon or Guardian, what are several other things you would like to see their IT groups do to assure the success of an IT project for your business unit? Defend your suggestions.