The Pressure Build At Big Blue

By Andrew Pollack
The New York Times

August 10, 1986

San Francisco -- On bulletin boards and in memos throughout the vast empire that is the International Business Machines Corporation, the word has gone out. Management wishes to reassure employees that they will not, repeat not, be laid off.

At a company in which a no-layoff policy has been as much a part of corporate culture as the blue suit and white shirt, such assurances would hardly seem necessary. But these are times of unusual duress for I.B.M. Even if the no-layoff policy survives, other aspects of the company's business seem about to change.

I.B.M., accustomed to nearly uninterrupted growth in profits, is likely to see its 1986 earnings drop for the second year in a row, the first time this has happened in a half century. Recently announced results for the second quarter show net income dropping 8 percent and profit margins are near their lowest level in years. I.B.M.'s stock has taken a tumble from its high this year of about $162 to about $130.

An I.B.M. that seemed all but invincible only a year and a half ago now has weak spots in its product line that have begun popping out like springs from an old couch. Sales of its new line of giant mainframe computers, which came out late in 1985, seemed to stall earlier this year, although I.B.M says sales are firm now and meeting expectations.

Its mid-sized computers, a series of machines that cannot talk to one another, are befuddling users and falling prey to a newly revitalized Digital Equipment Corporation. I.B.M.'s overwhelming dominance of the personal computer market is being sharply eroded by clones - less-complex, often non-brand-name machines produced at low cost in the Far East. And the new PC-RT, intended to spearhead I.B.M.'s entry into the fast-growing market for desktop engineering workstations, is widely viewed as anemic.

I.B.M.'s chief executive, John F. Akers, has attributed the company's woes largely to the weak economy and lackluster capital spending. Concerns that the investment tax credit will be removed by tax overhaul legislation have slowed purchases of multi-million-dollar mainframes. While analysts and outsiders say these factors are undoubtedly among the reasons for I.B.M.'s woes, they are not the only ones. Rather, they say, I.B.M.'s situation results from a number of tactical errors by management and changes in the fundamental economics of the computer industry that have hit I.B.M. all at once. And, many analysts say, the fundamentals indicate that I.B.M. might never return to the high profits and easy times of yesterday.

''They're frantically pushing all the buttons they pushed in the past, but this time, they're not working,'' said Brian Jeffery, research director at the International Technology Group, a consulting firm in Los Altos, Calif.

''I think we're entering a period where we are going to see an erosion of the huge gross margins we've seen from I.B.M. in the past,'' said Francis R. Gens, vice president of the International Data Corporation, a market research firm. Mr. Gens said what is happening in the personal computer arena might foreshadow a problem I.B.M. will face increasingly in the future: With many companies able to make hardware extremely cheaply, I.B.M. will have to make a choice between cutting prices to retain its market share -thereby reducing its traditionally high profit margins - or retaining profit margins but perhaps not growing as fast.

I.B.M., for its part, seems to want to retain both margins and market share. To do so, it is increasingly shifting emphasis from computer hardware to software, where profit margins remain high. It has invested tens of billions of dollars in plants and equipment for low-cost manufacturing. And now, perhaps belatedly, it is seeking to cut the costs of its 400,000-person workforce without any layoffs.

The company is consolidating operations in what appears to be a mystifying game of musical chairs. Manufacturing of printers and copiers is moving from Boulder, Colo., to other plants making such products in Charlotte, N.C. and Lexington, Ky. Meanwhile, 450 people in the Federal systems division, which develops systems for military and Government uses, are being moved from leased space near Los Angeles to Boulder, where I.B.M. owns the buildings. Manufacturing of the PC-XT and PC-AT in Austin, Tex., are being consolidated with those in Boca Raton, Fla., while manufacture of the Series 1 minicomputer is moving from Boca to Charlotte. While I.B.M. offers to move people or find new positions for those whose jobs are displaced, some outsiders and employees think one reason for the production shifts is to induce people to quit.

Elsewhere, I.B.M. is cutting outside consulting and travel expenses and bringing back work from contractors to give to its own employees. At the Endicott, N.Y. plant, workers have been cleaning and oiling their own machines since I.B.M. cut back on outside maintenance services.

The problems should not be exaggerated. I.B.M. remains the most profitable company in the world by far and, at least until the recent sour news, has been regarded as the world's best-managed company. It commands 40 percent of the computer industry's revenues and 70 percent of its profits.

Still, I.B.M.'s problems highlight the fact that even the best managers make mistakes. If they learn from them - and many outsiders are betting that executives at I.B.M.'s headquarters in Armonk, N.Y. will - the computer giant could emerge an even tougher competitor. In fact, some I.B.M. competitors fear deeply the struggle ahead to set earnings back on course at Big Blue. ''As I.B.M. tries to keep its head above financial water, in the process of flailing around doing that, a lot of people are going to be knocked off the raft,'' said John Cullinane, chairman of Cullinet Software, a competitor.

What follows is a list of the hurdles I.B.M. must overcome in the months ahead - and the strategies Mr. Akers & Company seem to be taking.

A shift to sales

When computers were behemoths with price tags to match, I.B.M. mainly rented its machines. But in the last few years it has shifted to selling them. Rentals, which accounted for 37 percent of revenues in 1981, accounted for just 8 percent in 1985.

But the shift, one of I.B.M.'s more fundamental changes, will make business tougher. Without dependable monthly income from rentals, I.B.M.'s financial results are more susceptible to economic cycles. Moreover, as leasing customers bought the machines they were renting, the shift produced a one-time spurt in revenues - transferring money that I.B.M. would have received over a long period of time into a quick injection of cash in the early 1980's.

''They padded their operations results with this one-time sale of assets,'' said Robert Djurdjevic, president of Annex Research, a Phoenix market research firm.

But padding results was not I.B.M.'s intention, analysts believe. As computers became cheaper, customers grew accustomed to buying them and I.B.M. simply followed the trend. Furthermore, in mainframes, in the 1970's I.B.M. faced competition from makers of compatible computers that would match I.B.M.'s design. To keep competitors at bay, I.B.M. began to introduce new generations of computers more quickly. By selling the machines rather than leasing, I.B.M. no longer has its own money tied up in the equipment and can afford to obsolete the old machines more quickly.

Still, the bolstering of I.B.M.'s results covered up, at least temporarily, problems that are starting to surface. And it indicates that I.B.M.'s hardware business has not been growing as fast as it seemed, meaning that growth is less likely to return to the fast pace of the early 1980's. If I.B.M. is to keep growing, it will have to sell more and more machines. And that is proving difficult.
A slowdown in growth

Charles C. Oldenburg, general manager of computer services at the Chevron Corporation, was incredulous. It was the fall of 1984 and a group of I.B.M. customers had just been briefed about I.B.M.'s growth projections. The company had said it would grow 14 percent a year for many years; such growth rates would have made it a $100 billion company by the 1990's and a $200 billion company by the turn of the century. ''I thought it was absurd,'' said Mr. Oldenburg. ''The growth rates just had to moderate.''

He was right. A survey of 500 I.B.M. users by the International Data Corporation showed that the main reason customers are not buying I.B.M. mainframes is that they simply do not need any new ones.

''In the 60's and 70's, we always needed more capacity than we could get,'' said Irwin J. Sitkin, vice president of corporate administration for the Aetna Life and Casualty Company. ''Suddenly that's stopped.'' Indeed, big companies have been increasing computer capacity at a range of about 30 percent to 50 percent a year since the late 1970's, and no longer need so much. At Aetna, for example, Mr. Sitkin said, the newest mainframe computer, an I.B.M. 3090 ''has more MIPs in one machine than we had in the whole shop in 1978.'' MIPs are millions of instructions per second, a common measure of computer power.

Moreover, for I.B.M. to keep growing, it must sell not only more computing power each year, but a lot more. As technology improves, computers become cheaper. The price per unit of computing power decreases by 20 percent or so each year, so I.B.M. must sell more than 20 percent more computing power each year just to keep revenues level.

So far, I.B.M. has been able to outrace the treadmill. Each price decline has spurred more than enough increases in demand to make revenues grow. But in the current downturn, demand does not appear to be as elastic as it had been earlier and price cuts have not produced the expected sales gains.

Many computer industry executives and customers think demand growth has stalled temporarily while vendors and users develop new uses for their machines.

At I.B.M., there are a number of new applications that could spur usage, but they are not quite ready yet. Networks are needed to link computers for electronic communications. Artificial intelligence, which would make computers easier to use and allow them to handle such things as drawing up insurance policies or offering financial advice, could lead to a new burst of computer use. I.B.M., which once shunned the field, is now getting involved in it. Image processing - moving images of documents around by computer - is another potential use. I.B.M. still has no product in this category, but customers say it is working on one. The company is also planning numerous enhancements to its mainframes.

But most of these improvements involve software, which takes a long time to develop. ''They know what the problem is but they don't have the solution,'' said Thomas J. Crotty, president of Gartner Securities, which closely monitors I.B.M.

A common language

Smaller computers pose a different threat to I.B.M. - that they will replace the large computers that account for much of I.B.M.'s revenues and profits. I.B.M. commands about 70 percent of the mainframe market worldwide, and its insulation from competition gives it gross profit margins of as much as 70 percent.

I.B.M. hopes that the proliferation of smaller computers will increase mainframe usage - that as more people work on computers at their desks, they will need to tap into the mainframe for data. But if that is to occur, they must be able to communicate with mainframes. Here, I.B.M. is weak. Its networks to connect computers are not ready. And its product line is a jumble of different computers, all incompatible.

The Digital Equipment Corporation, by contrast, has a single architecture - VAX - that extends from desktop computers to super minicomputers and can be easily interconnected. Digital, the No. 2 computer maker, jokes that the easiest way to connect two I.B.M. computers is to put a Digital machine between them.

The strength of Digital, whose earnings in the last quarter rose 138 percent, has so worried I.B.M. that it has recruited sales personnel in each branch office to learn about Digital's products and to concentrate solely on competing with Digital, according to former I.B.M. salesmen.

I.B.M.'s product jumble came from what looked like sensible management decisions made more than a decade ago. To keep nimble as it grew, I.B.M. decentralized, giving different divisions responsibility for their own products. Each division devised machines - the 8100, the Series 1, Series 3, System 36, System 38, Displaywriter word processor and then the personal computer. Now, however, I.B.M. is saddled not only with products that cannot easily interconnect, but also with the high costs of trying to upgrade and renovate many different product lines.

''The strategy of producing targeted products for market segments backfired,'' said Bob Evans, a former chief engineer at I.B.M. now with Hambrecht & Quist.

I.B.M. is moving toward product unification and networking, but it will never have a single product line like Digital. Most of the work is being done in the mid-range, where I.B.M. has said it will concentrate on the System 36 and System 38, which are incompatible but are being made more compatible. Rather than try to unify the hardware beyond that, I.B.M. says it will unify the software. In June, I.B.M. introduced numerous new products and improvements in its mid-range machines, which have been well-received by customers.

I.B.M. is also planning to move its mainframe architecture down to smaller machines. It is expected soon to introduce its smallest computer using mainframe architecture. With it, I.B.M. will be able to boast a line of computers, all using the same software, that run from small to large, rivaling Digital.
Beating back the clones

Computer hardware is fast becoming a commodity, something that sells on price and that any company can make. In retail personal computer sales, I.B.M.'s market share has slipped by 10 percentage points in the last year, and now accounts for less than half the I.B.M.-compatible computers sold. For many personal computer customers, there seems to be no compelling reason to buy I.B.M. anymore. In June, Mr. Akers told analysts that I.B.M. would consider withdrawing from part of the personal computer market if it became too commodity-like.

Such a statement might seem ironic. In the last decade I.B.M. invested billions of dollars to make itself the lowest-cost manufacturer, precisely because it saw that hardware was increasingly becoming a commodity that would sell on price. Yet even if it were to cut manufacturing costs to zero, it still must support a huge overhead in marketing and research. Moreover, its competitors accept lower profit margins to stay in business - something I.B.M. would prefer not to do.

''I.B.M. has thrown a lot of money against the wall to become the low-cost manufacturer,'' said Archie J. McGill, a former I.B.M. executive. ''The question is whether it was worth it.''

To compete on price, I.B.M. would have to cut profit margins. But instead, it wants to add proprietary value to its machines to distinguish them from the crowd.

As for the clones, in the short-term I.B.M. is trying to reduce their threat by bolstering relations with dealers. And it is thought to be preparing to introduce a new personal computer system that will sell for under $1,500 - a price that would make it more competitive with the clones.

In the longer run, analysts expect I.B.M. to introduce personal computers that contain more proprietary hardware and software. The new computers are expected to run programs written for the MS-DOS operating system that is now used on the current generation of IBM-compatible computers. But they might also include chips containing a second operating system proprietary to I.B.M.

These proprietary features are expected to begin appearing next spring in a new PC using the Intel 80286 microprocessor, the chip used in the PC-AT, but the new approach will not really get underway until late 1987 or early 1988 with machines using a new, even more powerful microprocessor, the Intel 80386. But the MS-DOS system has so much momentum now that I.B.M. will have to offer very special features to get people to buy the new machines and abandon the old standard.

Software and service

If there is one theme running through I.B.M.'s litany of challenges, it is that the company did not keep up with customer needs as well as it might have. In particular, it overemphasized hardware development and failed to provide software - to give users new uses for their computers, to tie computers into networks, to unify the product line, to add proprietary value to the personal computer.

Some customers say the sales force is also not as good as it once was. The I.B.M. sales force has always had the reputation of knowing more about a customer's business than the customer does. But some customers say that with the proliferation of products, the I.B.M. sales force has trouble keeping track of its own product line, let alone learning about competitive alternatives or a customer's needs. ''They don't understand Quaker very much,'' said Ronald T. Brzezinski, vice president of information services at Quaker Oats. ''A lot of them are not even aware of the alternatives available to a lot of us.''

I.B.M. is addressing these issues. Indeed, if a new I.B.M. is to emerge, it will be one focused far more heavily on providing software and service. Part of the reason is to spur usage and sales of computers. ''What you're seeing is a movement on a lot of different fronts to get rid of those things that are inhibiting our customers,'' said Victor Goldberg, an I.B.M. vice president involved with marketing.

But software and services - such as contract programming, consulting and maintenance - are also viewed as increasingly important businesses in their own right. Software and services grew 29 percent in the first half of 1986 over the same period in 1985. This category has risen from 18 percent of I.B.M. revenues in 1981 to 23 percent in 1985, and IBM expects it to rise to as much as one-third of revenues by the early 1990's. Software revenues alone have been growing more than 30 percent a year, and were $4.16 billion in 1985, or 8 percent of I.B.M.'s $50 billion in revenues.

In bidding on a job, I.B.M. says it is now willing to team up with other companies to provide a complete system for a customer, or to buy equipment and software from other companies and include it as part of its package if it does not have what the customer needs. ''In the past, we might have said 'We don't have one of those,' '' said Mr. Goldberg, the I.B.M. marketing official.

I.B.M.'s Federal Systems Division, which has long developed systems for government agencies, including ground control for space missions, recently also started to bid on commercial accounts. It will install the system and even run it if necessary - a service that companies such as Electronic Data Systems have long provided. It is building a travel agent office automation system for United Airlines and recently helped I.B.M. beat out Digital Equipment and Wang to be chosen as the sole provider of office automation equipment to Ford.

I.B.M. also said it will add 3,000 to 3,500 new people to its sales force this year, drawn mostly from the company and including many from its hefty middle management ranks. This move is designed to bolster the sales effort while reducing headcounts in middle management. I.B.M. is asking for volunteers, offering them the chance to move to another part of the country. ''Everybody's sitting around with brochures thinking about where they want to live,'' said one former salesman.

All these strategies will take time and are not without risk. The largest software companies have only $200 million in sales and it is unclear whether the software industry will be big enough to support I.B.M.'s growth to $100 billion or $200 billion in revenues. Further, notes Frederic G. Withington, an I.B.M-watcher at Arthur D. Little, I.B.M. might be stretching its programming talent too thin to provide programming for customers as well as for itself. Indeed, that is something I.B.M. is trying to avoid: At a time when the company is reducing its overall workforce, it is continuing to hire programmers and systems engineers.

IBM at a glance
All dollar amounts in thousands, except per share data

Three months ended
June 30 1986 1985
Revenues $12,268,000 $11,434,000
Net income 1,305,000 1,414,000
Earnings per share $2.12 $2.30
Year ended
Dec. 31 1985 1984
Revenues $50,056,000 $45,937,000
Net income 6,555,000 6,582,000
Earnings per share $10.67 $10.77


Total assets, Dec. 31, 1985  $52,634,000
Current assets  26,070,000
Current liabilities  11,433,000
Long-term debt  3,955,000
Book value per share, Dec. 31, 1985  $51.95
Stock price, Aug. 7, 1986
N.Y.S.E. consolidated close 131
Stock price, 52-week range 161 7/8-122 1/4
Employees, Dec. 31, 1985 405,535
Headquarters Armonk, N.Y.

GRAPHIC: Drawing; Photo of John F. Akers (NYT/Joyce Dopkeen) (Page 23); Graph of IBM's stock price vs. the Dow Jones industrials (Knight-Ridder Tradecenter) (Page 23); Graph of net income (Page 23); Graph of Revenues by business segment (Page 23); Graph of share of personal computer sales at retail outlets (InfoCorp) (Page 23)

Copyright 1986 The New York Times Company