IBM - Too Little Too Late

The Economist

December 14, 1991

Scenting extinction.

Mainframe computers made IBM. They may yet break it. By clinging to its beloved mainframe business, and the ways of selling computers that went with it, the biggest computer company of them all has jeopardised its future.

John Akers, IBM's chairman, has been half-heartedly trying to shake his firm out of its mainframe mould since 1988. To little effect: IBM's shares are at a nine-year low; its net profit for 1991 is forecast to be $2.4 billion, down 60% from 1990 and at a six-year low; revenues, forecast to be down by 5% at $65.6 billion, are contracting for the first time in 45 years; employees' morale is at a low to match. All this has at last persuaded the bureaucracy running IBM that changes proposed three years ago - to decentralise the company, slim it down and make it more entrepreneurial - must now be carried out.

Maybe. But the 1988 reorganisation was at the time described as every bit as epochal as the one announced on November 26th. It simply was not carried out. To judge by the stockmarket's reaction, investors believe this will be the case again. IBM's share price has fallen from a peak of $175 in 1987 to a recent low of $85, reached since November 26th. The shake-up announced on that day aims to recast IBM as a federation of flexible and competing subsidiaries. How that will happen is still not clear, either to outsiders or to employees.

What the market is saying is that gradualism is not enough. But analysts also fear that the drastic - and swift - changes needed to ensure the company's survival are beyond the abilities of IBM's management, all too deeply rooted in the firm's traditional blue-suited, white-shirted ways.

IBM's problems go beyond either the slump in the American computer industry that started in 1989 or IBM's famed inability to get innovative products to market quickly. What IBM is facing is a generational change in its industry. For a firm that owned the industry for 20 years, this demands a shift as wrenching as the one it made in the 1950s when mainframes supplanted tabulating machines.

The age of proprietary mainframe computing has passed, replaced by open-systems computing-networks of workstations and personal computers - unwittingly ushered in by IBM's own personal computer a decade ago. The old age was extremely lucrative for IBM. Gross margins on mainframes were upwards of 65%. Understandably, IBM - perhaps unconsciously - wanted the transition to modern times to be so gradual that it could milk its cash cow to the last drop. It was making its money from after-sales service of its proprietary mainframes, and doing it the creamy way by getting its service fees up-front in the purchase price of its systems.

When IBM controlled the systems technology, it could dictate the pace of technological transition because it dominated the markets for both services and engineering through the proprietary package that tied them together. Open systems, which allow all sorts of computers to collaborate, broke Big Blue's grip.

Many are betting that the company will never get it back. IBM's salesmen, say critics, still think in terms of big-time contracts and the fat margins that come with installing mega-machines for corporate customers. Its engineers are more excited about pushing the technological limits of switching devices, computer architecture and mass-storage equipment for large systems than they are about cramming computing power into cheaper and smaller boxes.

It is not that IBM misunderstands how the future is developing. It does have a toe-hold in technologies and businesses, such as its new high-speed RISC-based workstations, that are at the front lines of computing. It is pursuing joint ventures, such as those with Apple Computer, Motorola and Siemens, to take it where computing will go tomorrow, more swiftly, cheaply and safely than if it ventured out alone; it is also collaborating actively with original-equipment manufacturers (OEMs), such as Wang Laboratories and, in Japan, Mitsubishi Electric and NEC. (See box on next page for how IBM's Japanese operation is faring.)

Despite some spectacular flops, such as the OS/2 operating system for PCs, IBM is also pushing into software, particularly for service industries, where, Mr Akers reckons, the company's revenue is likely to grow faster than the industry average of 10-15%. What IBM does not have is a way to tie these various operations together in a way that will let it dominate markets as it used to.

As the boundaries between various sorts of computing blur, that may be impossible anyway. The most recently announced reorganisation seeks to give IBM's non-mainframe businesses more autonomy, in the hope that they will stumble into the right niches; it also aims to measure their success more by financial performance than sales volume. Yet the present mainframe-minded management also seems unable to let go fully of the old ways.

For example, while old product divisions will be turned into new autonomous subsidiaries, starting with printers and storage devices, these units will not have their own sales forces, but will rely on geographically organised companies for marketing and service support. And the target of 20,000 job cuts over the next year is one imposed from on high - not, as it should be, by the managers of each new subsidiary.

There is also a danger that IBM will end up creating a bunch of middling competitors in several markets. It has 80% of the $5 billion worldwide market for high-end disk-drives; but it has only a 20% share of the $53 billion market for all storage products, and less than 10% of the $30 billion printer market.

The more that software and systems-integration become the value-added end of the business, the less attractive it will be to customers to deal with a firm that is also a big manufacturer of its own hardware - and the more intense will be the conflict of interest in being one. Customers will take a lot of convincing that the various parts of IBM are truly independent. It may require the presence of substantial outside shareholders in those subsidiaries to dispel mistrust.

How fast Mr Akers can bring about change may depend on what happens to America's economy. At present he is relying on further (unspecified) cost-cutting to carry the company through a prolonged slump. But he also needs those savings to relieve pressure on the balance sheet caused by increased lending to customers - and does not have much margin of error if revenue does not pick up strongly. If all else fails, Mr Akers is only four years from retirement. IBM's best hope may lie in bringing in a strong outsider as his successor.