Outsourcing the IT Infrastructure

by Paul A. Strassmann


January 12, 2004

Information technology experts find it convenient to discuss the advances of computing in terms of technology "generations." There is a generally acknowledged evolution from the mainframe to the minicomputer, microcomputer, client/server and Internet stages of development. Yet defining IT strictly in terms of electronics doesn't necessarily offer an insight into economic gains.

Technology, in isolation, doesn't give us an understanding of the ideas that have propelled the progression from one stage of development to the next stage. For that, you need to consider the ideas of the 1991 Nobel Prize economist Ronald Coase, which are likely to guide the formulation of IT investments in the future.

The management of IT has always been steered by assumptions about economics. Grosch's Law, formulated by Herbert Grosch in 1953, stated that the power of computers grew as a square function of their costs. For almost 20 years, that simple conjecture (ultimately proven to be incorrect) dictated investments in corporate computing. Whenever possible, you upgraded to the largest mainframe computer you could afford. It was this compulsion that drove IBM to increase its manufacturing capacity for bigger mainframes until this strategy reached its economic limits and drove the company to the brink of bankruptcy.

Of all the computing "laws," I consider only Moore's Law to be based on verifiable evidence. It stated that the power of microprocessors would double every 18 months without corresponding increases in costs. The explosive proliferation of desktop (and laptop) computing propelled the practice of corporate IT for 20 years until it reached its current conditions, where the total costs of ownership have reached economic limits and the growth rate has leveled off, thus driving most suppliers out of business.

Robert Metcalfe's and George Gilder's formulations extended Moore's Law into the realm of communications. They stated that the value of interconnectivity would grow as a square of the number of connected devices, while the available bandwidth would expand much faster than the capacity of computing.

The multitrillion-dollar collapse of the technology bubble can be best explained as an unwarranted extrapolation of these concepts beyond any sustainable economic limits. Followers of Metcalfe's and Gilder's conjectures have induced the economic collapse of many communications companies. They bet shareholders' money on business that didn't materialize.

The problem with the Grosch, Moore, Metcalfe and Gilder formulations was that they concentrated only on the supply side of IT and paid little attention to the demand side. Those men were IT insiders who became wealthy from the rising prosperity for which they acted as prophets.

This is where a modestly compensated academic, Coase, comes in. He is the first economist of any consequence who has anything useful to say about information economics.

Coase studied why organizations are formed, what guides their growth and what leads to their demise. He observed that companies will expand until "the costs of organizing an extra transaction within the firm become equal to the costs of carrying out the same transaction on the open market." That is now known as Coase's Law. It represents an Information Age reformulation of the law of diminishing returns, which applied only to capital assets.

As organizations grow, they become complicated and find it costly to coordinate what they do. Coase observed that there are always companies that can deliver goods and services more economically than the dominant enterprises. If the more efficient companies become organized, they'll squeeze out those that have been unable to manage their resources. The only recourse for the inefficient companies is to shift inefficient functions to external suppliers. Thus, a carmaker will buy batteries from a supplier rather than manufacture them in-house if that's more cost-effective.

IT executives will have to accept that Coase's Law argues for outsourcing every IT function that can be delivered more efficiently by others. Corporate executives will demand that CIOs demonstrate how each element of their in-house IT spending has a lower cost than what's available in the marketplace. Farming out some of the labor to countries whose wages are a fraction of compensation in the U.S. is a relatively easy decision. The tough questions will concern whether to outsource a part or all of the company's computing infrastructure.

CIOs will have to justify contracting for the management of desktops, laptops or cell phones and having contractors provide security assurance, and they'll have to show that a commercially available standard application service can substitute for the legacy systems that have become unmanageable.

Acceptance of Coase's Law will force CIOs to prove that the total cost of retaining an extra transaction within the IT budget will be less than asking the competitive market to deliver it by means of network connectivity. It will convert every company from an organization-centric environment to a multisupplier services environment. This new focus will set the agenda for IT management for years to come.

Paul A. Strassmann (paul@strassmann.com) anticipates that within the next 20 years, most of the infrastructure supporting corporate computing will be purchased as a commercial service.

See more columns by Paul Strassmann

Copyright 2004 by IDG Communications, Inc., 500 Old Connecticut Path, Framingham, MA 01701.
Reprinted by permission of Computerworld

Go back up to the Strassmann, Inc. home page.