Out with conglomerates; in with alliances

10/01/1998

Out with conglomerates; in with alliances

Robert Haavind, Editor in Chief, [email protected]

Business structures are always changing. A few decades ago, huge conglomerates brought together disparate businesses into massive cash flow machines. Harold Geneen, who learned the style in the aerospace industry at Raytheon, honed it at ITT, where he became the king of the conglomerateers. His relentless push for growth built ITT into a multinational behemoth, and many other corporations followed the model. Geneen, however, was noted for the 16-hour days he spent learning more about the businesses than the managers who ran them, a regimen other top executives were unwilling to imitate. Conglomerates became financial holding companies rather than skilled operators of myriad businesses, often with poor results.

While the aerospace industry has continued in this mode, with some urging from the Pentagon (which now regrets the loss of competition), most of the business world has moved strongly away from this model. ITT, among many others, has spun off or sold subsidiaries that it had lapped up. Now corporate strategies concentrate on "core competencies," "doing what we do best," and the like. In this era, an acquisition binge is more commonly focused on a high growth market segment, such as telecom or networking, to build market share or add vital technology.

One reason for the change is that global competition is creating more demanding customers. They want more options but they also want quick delivery. Innovation must be coupled with agility and efficiency, particularly in high tech markets. It`s hard to do that in an accounting-driven business.

Outsourcing has become a popular tactic with firms seeking to shorten design and delivery cycles while concentrating on core competencies. Some process tool companies outsourced software for their tools, for example, and found out that this didn`t work very well if the tool developers and programmers were oceans apart. Strong design companies farm out their manufacturing to contract manufacturers. Fabless chip design companies tailor their designs exactly to what users need, and then have a foundry make the chips. But as we move into the System On Chip (SOC) era, foundries will need to become more versatile to provide special devices for low power, high frequency, mixed analog/digital circuitry designs, and also to provide diversified libraries of core functions that can be combined on one chip.

Rather than creating monster corporations through acquisitions, the new mode of business is alliances and partnerships, with each member contributing in areas where it offers special expertise. Contract manufacturers can buy the latest production systems and run them 24 hours a day no matter which vendors are winning in the marketplace.

The ultimate business form of the Information Age is the "virtual corporation," which builds value solely from intellectual property - knowledge and ideas. Rambus, for example, spent years doing the design for future ultra-high-speed memory-processor interfaces. The rationale was that licensing the Rambus designs would be much quicker and cheaper for chipmakers than designing difficult interfaces from scratch.

Will all this work better than having everyone working under one corporate umbrella? In theory it should. Each task will be done by those most expert at performing it. But things seldom work exactly the way economists say they should. What kind of commitment will there be in each sector of the supply chain? If things go wrong, whose fault will it be - who will sue whom? Will all the communications and automated systems link properly? Although the Rambus model may make sense, for example, chip firms are chaffing at having to pay licenses and some licensees are working on alternatives. Foundries make sense for high volume chips with moderate design goals, but will there be boutiques for the more specialized or higher performance devices?

Loose alliances may prove more successful than the full virtual corporation - it`s still too early to tell. But it sure will be interesting to find out.