The View from IBM

Lou Gerstner does have a vision: Network-centric computing. Here's an inside look at Big Blue's blueprint

By Ira Sager in Armonk, N.Y., with bureau reports
Business Week

October 30, 1995

In July, it looked as if Louis V. Gerstner Jr. had earned a well-deserved rest after two years at the helm of IBM. The giant was no longer listing, and cash was pouring in. On July 5, Big Blue closed the $3.5 billion deal to buy Lotus Development Corp., and the stock price was about double what it had been at the start of Gerstner's reign--making the CEO's options worth approximately $32.5 million. In August, Gerstner even took time for a five-day family vacation in Italy--sandwiched between calls on European customers.

Now, summer has turned to fall. And the outlook for IBM has turned from euphoric to cautious. On Oct. 11, Lotus chief Jim P. Manzi resigned, raising concerns over how the company is managing the strategically important merger. Manzi's departure is just the latest in a string of developments that are making this a tough season for Big Blue's CEO. In September, the company announced delays of two key products, and Jerome B. York, architect of Gerstner's cost-cutting crusade, quit. Both events raised questions about IBM's earnings momentum.

On Oct. 17, the company reported third-quarter earnings of $1.3 billion--excluding a one-time charge of $1.8 billion for the Lotus purchase--on revenue of $16.8 billion. The net was 6% below earlier projections. IBM's stock slid from a high of 114 5/8 in August, as some analysts cut earnings estimates for the year, and now trades in the mid-90s. As York told reporters when he left to join Kirk Kerkorian's Tracinda Corp., the cutbacks that helped boost 1994 earnings to nearly $3 billion--after losses of $16 billion over the previous three years--are largely over. "Management's attention is focused heavily on the revenue side of the equation," he said.

In other words, Job One is finding a long-term growth strategy for the world's largest computer company, a mammoth organization that has grown at only a 3.3% annual rate for the past decade. What Gerstner needs--and is quietly pulling together--is a plan to seize what may be IBM's last great opportunity: to become a leader in the dawning world of networked computing and electronic commerce. In this new wave of computing, communications rather than computation is the key. It involves everything from multimedia PCs to the vastness of the Internet.

"One of the great things about this industry is that every decade or so, you get a chance to redefine the playing field," Gerstner told BUSINESS WEEK in an exclusive interview. "We're in that phase of redefinition right now, and winners or losers are going to emerge from it."

The driving force is the arrival of low-cost digital networks that are spreading everywhere. Using high-speed fiber-optic cables and lightning-fast switches, digital networks will carry a whole new host of communications-based applications, such as teleradiology (transmission of X-ray images), distance-learning in the education market, or electronic transactions in business.

"Network-centric computing," as Gerstner calls it, is remaking the computer business the way the low-cost power of the microprocessor overwhelmed the mainframes and minicomputers of the 1980s. "Communications will change IBM more than even semiconductors have," predicts Howard Anderson, president of market researcher Yankee Group Inc.

It's an opportunity Gerstner says he knew he had to seize at the outset. But first he had more pressing chores, such as shoring up the balance sheet and restoring IBM's credibility with customers and investors. Also, he notes, there were problems getting some IBMers on board. "I sensed there were too many people inside IBM who wanted to fight the war we lost," he says.

Clear Signal. The lost crusade, of course, was the battle for supremacy in personal computers. With the IBM PC, the computer giant had a winner. But it lost its position by ceding technical leadership to Intel and Microsoft and the marketing edge to Compaq Computer and a host of cloners. Now, Gerstner has sent a clear signal within IBM to refocus on the future. He has disbanded a unit that was trying to create an alternative to Intel-based PCs that would have used the IBM-Motorola-Apple Computer PowerPC chip. He also has ordered a retreat from the costly losing battle that pitted IBM's OS/2 against Microsoft's Windows in desktop computers. The new contest is in networked software, Gerstner says--which is why he is willing to pay billions to acquire Lotus and its Notes "groupware" program (page 148).

The CEO still shies away from announcing a grand vision of IBM's future--after being lambasted for declaring early in his tenure that none was needed. But on a warm October morning, as he talks about the future of IBM and networked computing in a conference room next to his office, it's clear that he has one (page 152). Gerstner excitedly paints a picture of a 21st century business milieu in which networks become the lifeblood of corporations and the principle means of commerce. Eventually, he says, electronic commerce will extend to home shopping and other consumer transactions, too. By then, however, IBM may be reaching consumers with a network other than Prodigy Services Co., the long-struggling IBM-Sears Roebuck venture. "We are taking a hard look" at Prodigy, says Gerstner.

Gerstner's scenario hinges on some critical changes in IBM products and the company itself. If he can't get his troops to act quickly on his plan--and prove that IBM really has its networking act together--the computer giant's claim to leadership in the new era will remain wishful thinking. Competitors, naturally, say Big Blue will stay mired in the pre-network past. They point out, for example, that it has already missed hot markets for local-area networks (LANs) and related gear. "To say that IBM has seen the light--give me a break!" says Scott G. McNealy, CEO of Sun Microsystems Inc.. "We say to customers, IBM, DEC, and HP understand where you've been. We understand where you want to go with the Net."

What is IBM's network vision? Mainly, it's about bringing the company's traditional customers to network nirvana. "What is every major institution in the world focusing on?" asks Gerstner. "It's called reengineering. It's called getting competitive. It's called reducing cycle time and cost, flattening organizations, increasing customer responsiveness. All of these require a collaboration with the customer and with suppliers, and with vendors." All of these efforts, in short, depend on networking.

What Gerstner is doing--though so quietly that it's easy to overlook--is arraying resources from all over his empire for an assault on the wired world. And guess what? IBM has a lot to fight with. It has the world's largest private data network, an army of consultants and technicians to create custom networks and software for any type of business, and it even has some cutting-edge technology such as high-speed communications switches and software to enforce copyright protection on the Information Highway. Best of all, even IBM's biggest detractors concede that the company has enduring relationships with big corporate customers that are eager to squeeze time, cost, and inefficiency out of their operations.

IBM also has lots of competition. Everybody from AT&T and its long-distance rivals to all of the Baby Bells to Electronic Data Systems, Sun, Intel, and Microsoft to Oracle are racing into position. But Gerstner argues that his company has a nearly unique set of skills with which to bring corporations painlessly into networked computing. As the world gets wired, access to computing power and applications may become easier, but the complexity of building these systems--especially ones that operate around the world--soars. "IBM is one of the few companies that has the infrastructure to provide information systems on a global scale," says Gerald Ross, a co-founder of Change Lab International Inc., a Stamford (Conn.) consultant.

Big is Beautiful. Broadly, the plan is to use the computer giant's global reach, its expertise in dozens of technologies, and its knowledge of how major customers conduct their businesses to offer all sorts of computing resources across networks--either private or public ones. The network becomes an information-technology resource that customers can own outright or rent as needed. Someday, they might not even need a computer staff or data center. "People will just plug into the wall and not care where the computer is," says Harry Wallaesa, chief information officer for Campbell Soup Co., a major IBM customer.

In the wired future, says Gerstner, companies will be able to buy computing power and applications software the way they buy electric service. "They will rent them on a per-usage basis because they will be resident in the network," predicts Gerstner. "Think about the implications of a world in which the application is on the server--all applications."

Before that world takes shape, the 53-year-old CEO has a lot to accomplish. An important first step: convince programmers--in corporations and at software companies--to write applications for IBM networks. "They have to redefine where applications are developed, from the desktop to the network," says industry consultant Marc Schulman.

That's where Notes can play a critical role. It's already a big seller among the large corporations that use IBM mainframes. On Oct. 3, IBM announced Notes Public Networks, an agreement with 11 communications companies--including British Telecom, CompuServe, and Nippon Telegraph & Telephone--to carry the program on their networks. That will make it easier to experiment with the groupware setup. "You immediately lower the barriers for acceptance of Notes," says Fernand B. Sarrat, an IBM general manager for network services. "Customers can go directly into a Notes service."

Another way Gerstner is getting ready for the network-centric future is by building up Big Blue's outsourcing and systems-integration businesses. Outsourcing involves taking over a company's data processing operations and then selling back computing service or simply running the operation for a fixed annual fee. Campbell Soup, for instance, recently signed a $600 million deal for IBM to run its data center, network, and 7,000 desktop computers for 10 years. In the networked world, outsourcing will become more common and more sophisticated. IBM and other companies will design, create, and install "solutions" to business problems that customers can use by tapping into the Net.

Honor Code. Consider what's happening at DowBrands Inc. The maker of Ziplock bags and Fantastik is using an IBM network "solution" that it hopes will put it on the cutting edge of logistics. Every day, an IBM mainframe sitting in Rochester, N.Y., gets electronic updates from Dow's retailers on their inventories. The data are analyzed using Dow's production schedules, records of the retailers' past buying patterns, and other factors. When Dow taps in, the system can tell the company how much of each product to send to each retailer and when. Dow doesn't own the software or the computers. And it shares the service with rivals such as Procter & Gamble Co., which developed the software. Dow and the others trust IBM to keep the information secure and confidential, and each pays a monthly usage fee based on its volume of transactions.

Another convert to IBM's new pay-as-you-go system is Walgreen Co. The drugstore chain is testing a new system that allows doctors to electronically send prescriptions to pharmacies. The system saves time for doctors and drugstores, prevents pharmacists from misreading handwritten prescriptions, and can flag potential medical problems. Walgreen developed the software but sold it to IBM because it feared that doctors and other chains would balk at using a system run by one retailer. Now, other chains, such as Eckerd Corp., are trying it. Doctors get the service for free. Walgreen and other pharmacies pay a fee for each prescription.

Sure, it's penny-ante stuff now. But a few pennies here and there will add up. Walgreen and other drugstores will pay IBM from 12 cents to 18 cents per prescription. In the U.S., 2 billion prescriptions are filled every year, and IBM already has 25% of the nation's drugstores testing the system. What's more, once the system is in the doctor's office, other services can be sold--from patient billing to downloads of the latest articles from medical journals.

For customers, the savings are alluring. They don't have to buy multimillion-dollar mainframes, hire computer operators, or sink capital into data centers. "The economies involved in doing this are going to be significantly better than if you own your own computer," says Frank X. Dzubeck, president of market researcher Communications Network Architects Inc. Campbell Soup expects to save $200 million over the life of its deal.

IBM is already enjoying a boom in computer services. With $9.5 billion in revenues last year, IBM's outsourcing and systems integration business was second only to $10 billion Electronic Data Systems Corp. In the first half of this year, IBM's business grew at a 33% annual rate, while EDS is expected to grow 20% in 1995. On Oct. 11, Nynex Corp. picked IBM to help build a multimedia network and applications programs to sell on it. It's good business for Big Blue, says Ward MacKenzie, vice-president for Nynex Business Network Solutions. "In just my business market, we have access to 1 million customers," he says.

In addition to putting a new leg on the services business, Gerstner counts on networking to bolster other IBM divisions. Even as customers switch to buying computing across the network, they will still need new desktop computers, network servers, even mainframes in their various offices and plants. The IBM CEO is convinced this will ``reenergize'' the market for very powerful servers--which is how mainframe-class systems are referred to in the networked era. Conventional mainframes as well as IBM's new SP-2 parallel computers can act as transaction and database managers and keep secure huge quantities of data and applications software. First, Gerstner concedes, such machines ``have to be modernized and adapted very significantly.''

IBM may require an image overhaul, too. Rebuilding around a network-centric strategy is "exactly what they must do," says Robert Djurdjevic, president of consultant Annex Research Inc. "But that's not my perception of what they are doing." Gerstner acknowledges the image problem and says his style may be partly to blame. "I have this quaint view that you perform and then you talk about it," he says. "That's not the modus operandi of this industry--you talk about something, then you perform."

On Nov. 13, Gerstner will have his chance to talk. That's when he'll be the keynote speaker at Comdex, the computer industry's largest trade show. He'll use the speech to describe how network-centric computing is going to drive the next phase of growth for the computer industry and its customers.

Reining In. Meanwhile, he continues to put key pieces of his plan in place. Frustrated by slow progress in getting various IBM units to collaborate on network strategies, Gerstner is pushing executives to pull together technology and marketing efforts from various divisions. He has already pulled more than 100 far-flung network operations--including Advantis, the huge U.S. data and voice network jointly owned by IBM and Sears, Roebuck & Co.--into a single operation, called IBM Global Network. "Customers want to operate globally," says John M. Whiteside, who was recruited from MCI Communications Corp. to run IBM's network operation a year ago. "To do that, they need a single network." IBM Global Network is now one of the biggest Internet access providers, with more than 100,000 subscribers.

Gerstner is also making sure IBM develops content for the wired world. The Networked Application Services Div., headed by Sarrat, has $100 million to develop systems for electronic publishing, online catalogs, and videoconferencing that can be delivered over the Global Network, the Internet, or Prodigy. One new service, called infoMarket, uses a technique called "cryptolope" that lets publishers offer their wares in cyberspace without getting ripped off. If a consumer sees information listed online, he or she clicks a button, sending a message directly to the publisher, who then tells the client how much to pay to unlock the encrypted file. IBM, along with AT&T, Xerox, and a host of startups, is also working on methods for enforcing copyrights online.

If all of these efforts fail to make IBM a leader in network computing, what then? Then, the current concerns for IBM's growth potential could quickly turn to alarm as investors try to figure out where future earnings will come from. Until this summer, it looked as though Gerstner was on track to restore IBM's growth to the overall industry rate, around 10%, by 1996. He counted on winning back market share in key segments such as PCs and software. But IBM's PC business, though rebounding, is growing at below the industry rate, and software revenues are growing modestly. The services unit is booming, but mainframes--still the company's most profitable business--are flat. Because of delays in new mainframes and mainframe disk-storage products, nearly $500 million in revenue expected in the third quarter is being pushed into the fourth quarter. Gerstner also plans to take an $800 million charge in the fourth quarter to cover another spurt of cost-cutting.

And next year could be tricky. Should the U.S. economy slow and Big Blue's European business remain flat, some forecasters believe Gerstner may have to lower his sights to 6% or 7% growth for 1996--about what analysts expect for this year. That means IBM would continue losing share in key markets. Says a former senior IBM executive: "The more the company doesn't grow at the industry pace, the more it's behind the times."

Launching a successful network-computing strategy is no quick fix. It's bound to be a long and frustrating effort, full of false starts. Some IBM network projects already have fallen behind schedule or fallen apart. Intelligent Communications Services, which was hyped as a way to bring together mail from multiple electronic and voice mailboxes, was supposed to be ready by the end of 1994. But the group developing it failed to get a prototype to work and was quietly disbanded.

There's little time for such missteps, and Gerstner knows it. The Internet is sweeping the computing landscape, and if IBM isn't careful, it could overwhelm Big Blue's network efforts. "There's no question that the speed with which the Internet has emerged has caught all industries related to this technology by surprise," Gerstner concedes. Like other companies, IBM has seized on the Internet's World Wide Web as a nifty new way to communicate with customers, and it has a business, though small, in helping customers use the Net.

Parallel Computing. But the Web could also undo Gerstner's plans for Lotus. The Web setup can already provide some basic publishing and info-sharing tasks that companies buy Notes for. And in September, Web superstar startup Netscape Communications Corp. acquired Collabra, a software maker that competes with Lotus in groupware. "I think [IBM] bought Lotus Notes at the end of its life cycle--when everything is going Internet," says software magnate Lawrence J. Ellison, CEO of rival Oracle Corp. Just a year ago, Ellison himself considered acquiring Lotus. Now, he predicts: "Notes will be the Visicalc of groupware"--referring to the once dominant PC spreadsheet that was killed off by Lotus 1-2-3.

Gerstner has no choice but to get IBM zeroed in on the networked future. If the IBM net replaces--or even augments--the mainframe as the key information technology for big companies, IBM will retain "account control" and gain an advantage in selling a host of other products and services from PCs and network servers to huge collections of disk drives. "IBM is leveraging the network," says Jim Dowden, director of logistics for DowBrands. He's particularly excited about the new network-based services that require no up-front investment from him. "This baby is sitting there waiting for us," he says. "They're as close as the nearest telephone line." And one day, Gerstner hopes, network computing will be just as ubiquitous.

What's Down The Road

Network Application

An incubator for services and software designed to operate across a variety of networks, including IBM's own network, Prodigy, and the Internet. The 400-person unit oversees nascent efforts in electronic commerce, collaborative computing, content services, and interactive TV.

Rivals: Oracle, Microsoft, AT&T, Hewlett-Packard

Global Networking

With net connections in more than 100 countries and 850 cities, IBM is one of the biggest companies in networking. The heavy-duty corporate net serves 25,000 customers and doubles as the backbone for Prodigy. The network also delivers Notes-based applications and Internet access.

Rivals: AT&T, MCI, EDS, British Telecom

The Internet

The business of helping customers design and maintain Internet sites is small but growing. Customers from the Australian Chamber Orchestra to the Professional Baseball Organization of Japan have Internet home pages designed by IBM, which provides consulting services and security software.

Rivals: Sun Microsystems, AT&T, MCI, Digital Equipment

Industry Sales Units

IBM now reaches 80% of its customers through a sales force that is divided into 13 industry-specific units in such areas as financial services, travel, health care, utilities, and insurance. Such industry expertise should help the industry giant sell its products and services.

Rivals: EDS, Andersen Consulting

Information Services

One of IBM's fastest-growing units is Integrated Systems Solutions, which handles mega-outsourcing deals for the company's biggest clients. With its order backlog worth $30 billion, the organization is growing by more than 33% a year. Latest hot segment: Managing PCs and networks.

Rivals: AT&T, EDS, Computer Sciences

Lou Gerstner on Catching the Third Wave

IBM's CEO sees online computing driving all his businesses

Business Week

Louis V. Gerstner Jr. caused a stir two summers ago when he declared that the last thing IBM needed was to proclaim a grand vision. Gerstner proceeded to focus on cost-cutting and other management issues and proved the vision-hungry pundits wrong. But having studied IBM, the computer industry, and technology trends, the former McKinsey & Co. consultant has come up with, if not a vision, at least a plan for IBM. Relaxing in a conference room where he keeps a couple of computers to check out everything from kids' software to Web browsers, Gerstner shared his ideas with Information Processing Editor Ira Sager.

Q: You speak of ``network-centric'' computing. What is it?

A: The first wave of computing, 30 years ago, was driven by the technologies of host-based processors [mainframes] and storage devices. Twenty years later, we moved into a second wave, which was driven by microprocessors and simplified operating systems. The third wave of computing is being driven by very powerful networked technologies that provide very inexpensive and very wide [communications] bandwidth.

Q: What will happen to existing computers?

A: People have argued that the host-based [mainframe] model is dead. Well, it isn't dead. People are buying more of those units than they ever have. And no one should say that the stand-alone personal computer is going to go away. But there is no question that the PC-based model is now not the future. The future is a network-centric model. The focus moves to a network that draws out the best of both worlds. That model drives the strategy of this company. When people talk about IBM, it's interesting to hear them say: ``Well, I wonder if they'll get their new businesses to grow fast enough while their old businesses die.'' What they don't understand is, we are reconceptualizing our old businesses. We're bringing them into this new model.

Q: Will IBM put applications on a network and charge per-transaction or volume-based fees?

A: Absolutely. We call this business process outsourcing. For example, this Internet or networked world will put extreme pressure on traditional providers of financial services. We have no interest in going into the banking business. But we do have an interest in working with a series of banks in which we will provide networking capability through the IBM Global Network and networking applications that we will build. We'll take part of the transaction revenue. It could very well be on a per-click, per-communication unit. It is a strategy that brings our customers into this world of network services. Yes, we will be compensated in nontraditional ways.

Q: That provides a recurring source of revenue--as opposed to one-time product sales. Is this where you see network computing leading IBM?

A: That's the model IBM created with its mainframe software--monthly license revenue. We understand the concept of software paid for on a usage basis as opposed to a shrink-wrap, buy-it-once basis. Now, all of this is going to take time. It's going to evolve differently in different industries, in different companies, depending on their strategies. Large customers are going to take a very long time before they're going to rent mission-critical applications. People will not convert their own legacy systems to this. What will go on a rental model is the new stuff, like electronic commerce.

Q: In a wired world, must IBM remain a big maker of hardware and software?

A: People are going to need a lot of very powerful servers. They're going to need a lot of transaction software. They're going to need a lot of database software. They're going to need a lot of communications software. You don't necessarily have to provide all of these things, but we do. And as long as we're successful in each of them, we're going to continue to provide them. [If] we're not successful, we'll get out of the business.

Q: How does Prodigy fit into IBM's network-computing plans?

A: We are taking a hard look at our Prodigy strategy right now, along with [Prodigy partner] Sears. Short term, we're making a lot of changes. We brought in new management. They've done some new things with their system. Certainly their merchandising and marketing has gotten better. But the long-term strategy for IBM and Prodigy is not clear. We will sort that out in a relatively short time. The fundamental issue is: In the world of the Internet, is there a place for a packager of services? Does the customer want to go surf the Net and go to every one of 50,000 Web sites? Or will people pay a reasonable amount for somebody to go out and preselect and package what they want? My guess is they will both coexist.