email article printer friendly

Solving Real-World IT Architecture Issues: The Edge of Chaos
Examine how real-world companies solved their IT application architecture complexities.
by Lauren Y. Dresnick

Enterprise Architect Summit, May 2006

 Listen to the audio and watch the slides! (Running time: 1 hour)

Chris Barlow Consultant, McKinsey & Company

Chris Barlow began his Enterprise Architect Summit keynote by asking the audience, "Have any of you heard of my company, McKinsey & Company?" Before attendees raised their hands, Barlow said that The Wall Street Journal described McKinsey simply as a "secretive company that works with CEOs."

He provided more detail: "We work on strategy and operations with CEOs [and CIOs] from The Global 500, the world's biggest companies." In his seven years at McKinsey, Barlow explained, he has gotten "a pretty broad view of some of the biggest businesses in the world, and some of the problems they face when they're trying to move ahead on architecture." These experiences with McKinsey's clients comprised the core of the presentation.

Barlow described his work further as "walking on the edge of chaos," a theme he drew on throughout the session. "Most of the CIOs I work with are really struggling with architecture issues," he said. "They're struggling with their current application portfolio, data structures, and infrastructures," trying to get to something that will drive impact through the company, he said. He compared his experiences to those of an ER doctor who thinks that everyone in the world is a gunshot victim, "because that's all they see day after day. They never see healthy people." Barlow doesn't see many healthy organizations.

The primary problem for most of Barlow's clients, he said, is their legacy architectures. So McKinsey developed Principles of High-Performance IT to help their clients achieve business impact with IT efficiency:

• Make IT a business-driven line activity, not a tech-driven staff function.
• Make IT funding decisions like other business descisions—on the basis of value.
• Drive simplicity and flexibility throughout the tech environment.
• Demand near-term, business results from development efforts.
• Drive constant year-to-year operational productivity improvements.
• Build a business-smart IT organization and IT-smart business organization.

Real-World IT Architecture Problems
Barlow described some of his client's biggest problems. Many CIOs and CEOs are under tremendous pressure to cut their IT spending, "to get by with less ... to take [their] budgets and cut millions and millions of dollars out of it." Or, his clients are trying to balance increasing production costs while still delivering the same high-quality software. But "legacy architecture is expensive," he said.

Barlow used a retail client to demonstrate how some organizations cope with their legacy architectures. This company had a customer database that housed information in 40 different locations, and each location operated on various technologies. "[They] had about a thousand different programs accessing [their] customer information with about 16,000 access paths," he said. Plus, the company had an increasingly constrained IT budget with which their customer database had to be maintained. Because of these problems, the organization's IT department had to say, "Sorry, we can't change that customer information. We can't ... make a change to the database that's going to enable some kind of business strategy," Barlow said.

A similar example of how architecture constrains business strategy can be drawn from the banking industry. "Most banks are moving to a multichannel [management] strategy," he said. This management strategy allows you to access your banking information at a branch location, through a call center, or over the Internet. "Being able to work across [these] channels is critical for banks," he said, because it helps deliver better service and capture more customers.

However, these channels often operate on separate systems in a legacy banking architecture. The IT reality is "redundant and inconsistent databases in each channel," Barlow said. "They synchronize [these] back-end databases periodically, but there are still latency issues, where the data is not the same between the systems, and the look-and-feel of the systems is different." Barlow identified some of the problems that arise from these latency issues: You have limited cross-selling opportunities, substandard service quality, and constrained customer behavior data.