IBM Rethinks

New Importance Placed On What The Customer Wants

The Economist

November 17, 1990

With sales five times those of its nearest rival, IBM towers above the rest of the computer industry. Each year it spends more on research and development than any four other computer companies put together. It has the most salesmen and technicians around the world. It has the broadest product range. In short, by any measure IBM is the biggest. The challenge now facing it is to make biggest best.

To do that, IBM is trying to give a new meaning to the phrase "economies of scale". Instead of the old assembly-line gains, it is looking for economies of scale in information-gathering. Because IBM sells the most things to the most people, it should have the most information about what computer-buyers want. The difficulty is to make that information accessible throughout the company, and to act on it.

IBM could use such a competitive fillip. Its market share has shrunk with each succeeding generation of computer technology. According to Dataquest, a firm of computer-market watchers, IBM holds half the world market in mainframes, which were state-of-the-art in the 1960s. In minicomputers - the smaller, more flexible computers created by companies like Digital Equipment in the 1970s - it holds about 15% of the market. In personal computers it holds only just over a tenth.

In the mid-1980s, IBM executives were still confident enough to predict that their company would at least match the average 15%-a-year growth of its markets. They were wrong. IBM's sales grew at less than 6% a year from 1985 to 1989, to $63 billion. In some products, particularly software, it was out-innovated by young companies started by blue-jeaned entrepreneurs. In others it was out-manufactured by the Japanese. And in some markets IBM just tripped itself up.

In 1987 IBM's chairman, Mr John Akers, set up meetings between his top managers and his biggest customers. The results were not flattering. Customers lambasted IBM for arrogance - its unwillingness to listen and inability to adapt to customers' needs. Today IBM remains unloved by many of those who could do most to help it, notably the systems integrators who install many corporate computer systems. Instead of sweating blood to help customers, runs the most common complaint, IBM tries to bully them into doing things the Big Blue way.

Such complaints go to the heart of IBM's problems. To bully customers is to risk losing them - as IBM sometimes has. In the old days, each maker's computers worked only with other machines from the same maker. Users might be fed up with their computer-supplier, but the alternative was to tear out the whole system and start afresh. Today, the emergence of "open systems", based on non-proprietary technical standards, means that many different brands can work together. Customers are much freer to mix and match suppliers as they see fit.

That is why IBM now feels the need to become more responsive. Greater responsiveness would remove its customers' biggest gripe - and give it an advantage over the Japanese, who can match many of its economies of scale in research and production but (so far) lack both deep knowledge of western computer-buyers and the relationships from which such knowledge springs. It would also give IBM an edge on the small, entrepreneurial makers. They have market knowledge, and are unequalled in their ability to innovate in response, but they lack IBM's technological and manufacturing resources.

The snag is to put IBM's strategy into effect. It must change the attitudes of its 370,000 employees. That has begun. Mr Larry Ford, in charge of IBM's internal automation, has spent nearly two years laying the foundations for vast databases to hold and distribute information about IBM's customers and products. To encourage employees to gather and use such information, IBM's president, Mr Jack Kuehler, has been preaching a simple message:

Get out of your offices. Find out what you're doing right. Find out what you're doing wrong. Do something about it.

At the most recent of IBM's yearly gatherings of top management, Mr Kuehler reinforced those words with a new quality programme. Each of IBM's business processes - billing, order entry, even public relations - is now being treated as a product in its own right. Everybody's performance, from top to bottom, is measured in terms of quality and customer satisfaction.

If any company has the managerial skills to create a responsive giant, it is probably IBM. But responsive and giant could yet prove to be contradictory words. One reason that small companies pay so much attention to customers is that they do not have much else to concentrate on. Big companies do: they are held together by layers of management, a structure independent - only too independent, sometimes - of the market. Information does not diffuse readily through layers of management. Yet without them any big company risks breaking up.

That is exactly what some experts reckon big companies should risk. Understandably, IBM's bosses do not agree. They reckon they can marry size and responsiveness. To see how it is working, look in turn at each of the three pillars of IBM's strength: products, marketing and management.

"Do-it-yourself" might be IBM's technological motto. It produces in-house about four-fifths of what it sells. Though some Japanese companies are just as vertically integrated, most western rivals prefer to buy in parts from outside. Compaq, number two to IBM in personal computers, makes fewer than 20% of the parts in its machines. It prefers to focus on design, manufacturing only the key components for itself.

With one or two plants specialised in making chips, disk drives or what-have-you for each of its three global regions - America, Europe and Japan - IBM can maintain in-house production at high levels. It can invest long-term, averaging out prices and income over the market's peaks and troughs. Besides, argues Mr Kuehler, vertical integration aids product development. Much innovation involves taking big, expensive components from big, expensive computers and making them small and cheap enough for a desktop. So the high end of IBM's product range can give it a glimpse into the future of the low end, and familiarity with the whole spectrum of technology can speed efforts to bring that future about.

But these advantages are paid for in flexibility, managerial strain and time. To its credit, IBM has - in most products, most of the time - kept itself at the technological forefront. Instead of competing in components, it compares its wares to those of rivals and improves them where necessary - something that many other vertically integrated firms have not managed to do. But cost and quality, the twin goals of IBM's manufacturing investment in the 1980s, are not enough.

Mr Pat Toole, the man in charge of IBM's manufacturing strategy, recently added a third item to the top of his list of priorities: time to market. IBM can no longer afford to take five or six years shepherding products from laboratory to marketplace. So Mr Toole is trying to get all the parts of IBM working together on new products from the start, especially in manufacturing and marketing. That way development schedules can be shortened, and IBM can avoid delays that arise when designs have to go back to the drawing board because they omit customers' favourite features or are too tricky to build.

Mr Toole is building new project-management systems which help to co-ordinate the flow of information between the various groups involved in each project. There are signs of improvement. IBM's recently introduced PS1 home computer took 15 months from drawing board to launch. But the company needs to do better still. To make things harder, the variety and quantity of information that it must take into account are exploding. A wider and swifter spread of information within the company is only part of the solution. IBM must also swap more technologies with the outside world.

That is not easy either. Computer-users have spent 20 years collecting whatever would do the job at hand: a minicomputer to balance a division's books, a PC on a salesman's desk to analyse market trends. Now they want to link all those machines into seamless corporate networks. They know this integration requires standards and discipline. But they also want to keep their favourite goodies from the magpie years, no matter what brand-name they bear.

To meet these conflicting demands, computer-makers must share technology much more today than in the past. IBM, for its part, is packaging its vision of integrated computing in two forms. For buyers it offers OfficeVision, a collection of (mostly IBM-brand) products capable of sharing information across an organisation. For borrowers, integrators and (shudder) imitators, IBM is also publishing its Systems Application Architecture, the set of technical standards and disciplines on which OfficeVision is based. Both include many ideas and technologies from outside IBM, and development of both is running behind schedule. That is probably not a coincidence.

IBM manages its technological horse-trading mostly through alliances. It offers small companies access to its vast marketing network; they adapt their bright ideas to IBM's technological environment. In software, IBM now has hundreds of alliances. Unfortunately, managing alliances is messy, and it can wreak havoc with the disciplines of IBM's vertically integrated production.

Take, for example, a recent upset in what had seemed IBM's cornerstone alliance: that with Microsoft, the world's biggest supplier of software for personal computers. Microsoft created the DOS operating system for IBM's best-selling personal computer. It worked with IBM to create DOS's planned successor, OS2. OS/2 plays a key part in IBM's vision of the future of computing: an "information utility" that can hide all the technological complexity of integrated computer networks behind a single plug in the wall, giving the user access to whatever information he may want without the need to worry about where that information is stored or computed. OS/2 was designed to offer painless access to the world behind the plug.

The trouble is that customers have been much slower to buy OS2 than either company forecast. Microsoft recently lost patience, and, to IBM's annoyance, switched its attention to a product called Windows. Windows does only part of the job that OS/2 can do. But this seems to be the part that customers now want: the latest version of Windows has been selling like hot cakes.

For Microsoft, this switch is consistent with its mission: providing the personal-computer software that people want and will pay for. IBM's grander vision, by contrast, leaves it with less freedom to change plans. What for Microsoft is a market opportunity to be seized is for IBM - despite its zeal for responsiveness - a complication to its vision of computing.

Happily for IBM, many customers do not care much about technology: they will let IBM choose that, so long as its choices do the job that they choose to get done. That is where IBM has fallen down in the past. So now it is trying to shift the centre of its marketing people's interest away from itself towards their customers.

Until recently, IBM's sales force was organised by product and by country, like the rest of the company. This meant that - despite the company's size - it could be difficult to put together systems that combined IBM products from different countries or product groups. Multinational companies complained that terms of business, and standards of service, differed annoyingly from country to country. Transplanted Japanese executives lamented that they could not get Kanji (Chinese-character) word-processors for their American offices.

Now IBM is organising its sales force by customer or, for smaller customers, by category - shoe retailers, say. It will be up to the salesman assigned to Ford Motor or to shoe-shop companies to get the rest of IBM to do what keeps his customers happy.

The first beneficiaries of this shift have been IBM's multinational customers. They account for a quarter of its sales. Yet until about a year ago they had to deal, separately, with local IBM offices in each country where they did business. Today they have the option of one-stop shopping through a global marketing programme. Four-fifths of them have jumped at it.

Each global account is the ultimate responsibility of a single IBM salesman, backed by an international team. To find out what a customer wants, salesmen from around the world are typically flown to its headquarters - not IBM's - once a year for several days of meetings. Terms of business are global, and the customer can shop from IBM's worldwide product catalogue.

It will be harder to extend this focus on the customer to smaller firms. Mr David McKinney, boss of IBM's European operations, is trying to break down some of the national boundaries in his domain, and to replace them with a concern for specific industries. Thus IBM's British arm is being built up as a centre of expertise in European financial industries. Manufacturers are the job of IBM Germany. France is to specialise in telecoms. And the Swiss, true to stereotype, have been told to supervise quality.

Meanwhile, IBM is encouraging its salesmen to identify their own opportunities for serving customers, particularly small, fast-growing ones. A manager who has identified a new market - say drive-in photo-processors - can try to get backing to create a package of IBM products and services for them. If he thinks it appropriate, the "opportunity manager" can even organise business alliances to bring in skills that IBM lacks. That, at least, is the theory.

All of IBM's marketing intelligence is being recorded in vast and growing databases, which are gradually being made available throughout the company. To provide the managers to scour these databases in search of opportunity, Mr Walt Burdick, IBM's head of personnel, is redrawing his picture of the ideal IBMer. IBM, he says, has always looked for people who are "achievement-oriented". Traditionally, that has meant eagerness to get ahead, to climb a hierarchy. Now the company is increasingly seeking people for whom achievement means learning new skills or tackling new challenges.

Mr Burdick hopes bright people will come to IBM because it is an interesting and pleasant place to work. The company has always cherished these qualities. It spends heavily on training, giving employees opportunities to learn everything from computer-programming to the management of people. Unlike some companies that claim to, it really cares about its employees (indeed it can be almost humorously paternalistic: not only are retiring executives given $5,000 to learn a hobby, but those slow to spend it receive a series of increasingly concerned letters about their leisure activities).

For all its good intentions, though, IBM lacks the free-wheeling openness of many other computer companies. It is secretive. At a recent management gathering, one executive referred to Motorola to illustrate a point. The slide showing Motorola's published annual results was marked "IBM confidential". More important, IBM remains highly centralised. Though it is now trying to push authority down the hierarchy, big decisions are still taken either by the four-member management committee or by one of 65 or so corporate officers. All these executives have been with IBM for at least 25 years. This breeds cohesiveness and a clear sense of vision and culture. Ask IBM's president, Mr Kuehler, how he thinks of the company and he begins: "Well, we've all worked together for years..." Some people would not think that an advantage. Cohesion does not always breed ideas.

In IBM's vision of the future, then, self-motivated employees use their initiative to spot and tap new customer needs in fast-changing markets, aided by an inflow of technical and market knowledge. It is information that holds this vision together. Information about customers' wants guides product development and marketing, and each sale gathers more information to improve the accuracy of that guidance system.

Information, though, must flow out as well as in. Computer technology changes fast, and customers may not know what they want until they have an idea of what will be on offer. IBM traditionally has been less than forthcoming about its plans, even to its own people. Its management disciplines might prevent a frank exchange of views. It has sometimes been overprotective of its technology, stamping on small outside companies that had taken a good IBM idea and made it better.

Perhaps surprisingly, the best hope - for the moment, anyway - of keeping information flowing may lie in Mr Kuehler's quality programme. Quality cannot be achieved by decree: all employees, high and low, must work together to define and solve problems, which means they must engage in exactly the sorts of argument and discussion that IBM now needs to promote.