Next, a sound core or context analysis can form the basis for such a transition, helping the organization set priorities by their contribution to strategic differentiation and management of mission-critical risk. "Core" so distinguishes a company's offerings from competitors that they make those competitors unappealing. "Context" makes business processes more productive. Diverting efforts from non-core mission-critical work that drains resources to little purpose to either will accrue real benefits.

"There's a tendency for mission-critical work to accumulate lots of resources because people don't want to get punished," Moore says. "But those are the same resources that you need to deploy the next generation of innovation to get upside."

Pay for innovation by extracting resources out of context to repurpose for core. This funding strategy has the added benefit of reducing inertia, which builds up in context activities. Then focus resource management on balancing the disciplines of invention, use and optimization.

"Most people think established enterprises lack invention capability, but in fact they invent plentifully but have trouble deploying their inventions. The reason is their deployment resources are stuck trying to grind out the next quarter's numbers from an increasingly ageing prior set of innovations. If these companies could better optimize their resource utilization in established markets, they could free up the necessary resources to build the market for their next set of inventions," he says.

There is also typically intense competition for resources, and high levels of resistance to change coming from resources dedicated to maintaining existing processes.

"If you look at the way companies budget, they usually budget by saying: 'Well, let's mark up last year's budget and create next year's budget'. Well what that effectively does is it gives everybody who is legacy the first dibs. As a result, we have these processes which cause us to create more and more inertial momentum around what we did well in the past," Moore says.

Another prescription for change is resource recycling, a program for retaining a workforce while actively using outsourcing. Resource recycling identifies three key roles in the life cycle of an innovation: invention, deployment and optimization. Once the innovative process stops being core and has become context, optimization's goal is to prepare it for outsourcing. There is an enormous backlog of this optimization work to do, so when companies do outsource, they should recycle the optimizers to work on the next batch of context. This frees "deployers" stuck on undifferentiated work for reassignment to the next wave of innovation. That in turn frees inventors to go off and invent "the next new thing". In this model, more of the workforce stays with the institution whenever legacy work is outsourced to another company.

"Part of what we have learned over the last five years is, you cannot ask part of the organization to sacrifice for the other parts," Moore says. "Every part of the organization has to go through this renewal process all the time. Otherwise you have the cash cows that are resentfully and sullenly giving milk to a bunch of heifers who think they can't do anything wrong, and the entire corporation gets pretty dysfunctional around those issues.

"So we're trying to help people say: 'No, recycling is just part of the normal course of business', and making it a normal part of the business is part of our strategy for getting people to adopt behaviour which in the past they have not adopted," Moore says.

The same thinking should apply to use of offshoring and outsourcing as tools.

Offshoring taps into low-cost labour markets to enable lower price points in commoditized offerings. A high-standard-of-living economy's best defence against offshoring is to differentiate in ways that create customer preference over the commoditized alternative. To differentiate this forcefully, companies must focus ever more of their resources on a particular innovation, Moore says. This is where outsourcing comes in. The formula is to outsource your context so you can focus on your core. There are plenty of challenges to delivering this formula, but since there are few viable strategic alternatives, these are challenges we must master to continue our current standard of living, he says.

"I'll give you a classic example of something that you would think would be outsourced by every corporation in the world and it's probably outsourced by virtually none, and that's e-mail. E-mail is enormously burdensome to manage, at least it can be, and there's very little competitive advantage to an e-mail system. So why in the world are you spending in-house resources there?

"The irony is, because you're spending your in-house resources there, then when the big innovation project comes along, you don't have anybody in-house to do it, so you end up taking it to a consultant. So now you've taken the thing that is unique to a consultant. This is bad for all kinds of reasons, not the least of which is, once they're finished with your project they're going to take that knowledge and sell it to a competitor. You're guaranteeing that you're going to have a short life cycle to whatever it is you are trying to do differently the next time around.

"So we've got it exactly ass-backwards, exactly 180-degree reversed," Moore says. "It's not that IT should not outsource, it is that it should outsource context so it can insource core, and the irony is it is doing just the opposite."

No self-respecting giraffe would ever make the same mistake.

SIDEBAR: When SOAs Rule the World

Author Geoffrey Moore explains how SOAs will become the underlying force of the new data centre

By Julie Bort

Geoffrey Moore, known for his IT marketing book, Crossing the Chasm and more recently Dealing with Darwin: How Great Companies Innovate at Every Phase of Their Evolution, is widely regarded as an industry visionary. Lately, he has become one of a handful of clear voices that help corporations and vendors redefine their IT infrastructure goals based on new data centre technologies. Because the global extended enterprise - the "inter-enterprise" - will be the business model of the future, network executives should be crafting long-term architecture plans to ready themselves for it, says Moore, who is managing director of consulting firm TCG Advisors, founder of the Chasm Group and a venture partner at Mohr Davidow Ventures.

TCG Advisors envisions a new computing model for when the inter-enterprise becomes the basis of all IT infrastructure. Where can we expect the most change?

All the action is going to be in the middle [layer] in preparation for a significant change at the top in about five years. We are all [preparing] for a new business model: the inter-enterprise network value chain. For the inter-enterprise network value chain, traditional business applications need to be re-architected so they can cross company boundaries. When people look at this re-architecting, they see two huge barriers: the middleware, because our software doesn't work this way, and the business process layer, because people don't have a lot of experience [with it]. There will be a fair amount of trial and error before we figure this out. Even thought leaders are struggling.

Tell me about the middleware barrier - that our software doesn't work this way - isn't that the reason for service-oriented architectures (SOA)?

Absolutely. [Enterprise] software is set up to work on a client/server structure in a single corporation dividing the presentation layer from the compute and data layers. The goal has been managing information. We're shifting from managing information to managing processes. Information is an important attribute in process management, but it's not the goal. So trying to turn data into information is the wrong way to look at the problem. What we are trying now is to manage processes across multiple states, where any portion of the process can be in multiple states. So you have to keep state - which is a computing idea - and you have to coordinate actions among self-managing logic. That's the service-oriented architecture paradigm.

The concept behind SOA isn't new. What has changed to make it more valid this time?

Globally, the low-cost economy of Asia is now coming online and Asian vendors are a resource to the highly developed economies for cost reduction. So, on the one hand, you want to reach out to this system to lower your own [costs]. But the very same vendors are going straight to your customers and so they are commoditizing industries that for a long time had high margins. These include the PC industry - witness IBM exiting to Lenovo - consumer electronics, appliances, furniture and apparel. Those industries now have to reinvent themselves. This architecture is going to be necessary to recombine the world into something newly valuable that hasn't been done yet. What service-oriented architectures let you do is recombine - they are like Legos - to make all kinds of innovations out of the existing components of the world. This is very productive. If you're stuck in a client/server system you can't participate in that.

For example, the traditional approach for manufacturers is to design, source, build and sell. But in a SOA world, a company might start with sourcing: find out which are the lowest-cost parts it can source, figure out what it can design with those parts, sell the design ideas and only build what customers will buy.

How does a SOA affect the rest of the stack?

If you implemented just in software, you would bring the entire world to its knees. Multiple revolutions of semiconductor chip redesigns will be required to handle service-oriented processes. The classic microprocessor architecture that Intel dominated is going to become much less relevant. More relevant will be a network processing style of architecture. Everything is going to become a router. Router chip ASICs are going to become more central to the design/architecture of systems than the compute cycles of a microprocessor. This means that a company like Cisco is going to run head-on to a company like Sun because both are going to want to build the machines that handle all this.

What should network executives be doing now to ensure they are building a viable architecture for the new data centre?

It behoves them to have a business architecture that would imply a software and a hardware architecture underneath it. And it's a 10-year [plan] - this isn't going to be a short-term thing.

What industries are most capable of inter-enterprise architectures on a large scale?

Value chain re-engineering is about doing things hundreds of thousands and even millions of times. Probably the automotive industry is [closest because it] has a highly distributed value chain. What's interesting - and this is true of all inter-enterprise success - is that a highly concentrated part of the value chain forces everybody else to dance to its tune. You see this on the supply side in the automobile industry and on the distribution side in the retail industry [with companies such as] Wal-Mart.