Interview: IBM CIO Phil ThompsonBy CIOinsight
IBM began wrestling with its complexity demons a decade ago, when Lou Gerstner took over as CEO. Since then, IBM has cut IT spending by 31 percent. CIO Insight talked with IBM CIO Phil Thompson about the roots of complexity and how IBM has tamed the beast.
CIO INSIGHT: What's the root cause of complexity?
THOMPSON: It started with the concepts of management back in the early 1900s. Large companies learned from Ford Motor Co. the idea of specialization of task so they could keep people highly efficient, productive and focused. But things became less integrated and more fragmented. Instead of having one finance person, you had a person who handled accounting on the ledger, you had another person who handled planning, you had another person who handled treasury issues.
What we're talking about is dicing up business processes. At first you make just one black Model T, then you start throwing in power steering and different colors and different kinds of seats. Our business is the same way. Now you have all flavors of PCs, all kinds of software. You have low-end servers, high-end servers, storage area networks. And each one has an organization and a business model about how they go to market, the channels they use, the pricing methodologies and how they're going to support it after sales.
IBM was in trouble in the early 90s. To what extent was that a complexity problem?
Lou Gerstner would say that IBM losing its way with regard to the customer was the first and foremost thing. And, second, he would tell you that the integration of IBM as a company was the biggest value we could bring to customers. If complexity isn't number one, it's number two.
When Gerstner came on board, we had 24 different business units and we were not sharing services across the company; every unit had its full complement of everything. But we've gone from 55 data centers down to 12. We had 31 different network providers; we're down to one. We had over 100 CIOs; today we have one. We had over 100 different desktop configurations; today we have four.
Looking back a decade, did complexity hurt IBM in reaching its business goals?
Is it true today?
It's less obvious, but there's still tremendous opportunity for refinement in how we satisfy customers. We've worked our way through most of these problems and achieved a significant amount of integration end-to-end. But complexity also comes when you just have too many of everythingtoo many data centers, too many commerce engines, too many network providers, too many client desktop configurations.
Do you ever really solve the complexity problem?
I don't think so. To think anyone would ever have that one big thing in the sky that's fully automated and sort of like Star Trekthat isn't going to happen tomorrow.