IBM Shares Its Secrets
Crunched by low-cost competition, Big Blue is betting that giveaways of precious technology will expand the market--and boost its own prospects.

(FORTUNE Magazine) – It's IBM's nightmare. In a conference room in Bangalore, a team of retail experts at software company Wipro are redesigning the consumer experience for a major U.S. retail chain. They're methodically evaluating the checkout area. The client wants its processes to be state of the art, and Srikant Shankaranarayana, Wipro's brainy, intense 44-year-old general manager for retail solutions, is pushing his consultants and engineers to ask tough questions: Should salesclerks carry handheld transaction devices or stand at cash registers? Which merchandise should be tracked electronically? How much information needs to be in the database to ensure that discount promotions don't last longer than necessary?

Those are exactly the kinds of questions that IBM wants to be asking on behalf of retail customers. And the fact that such companies as Louis Vuitton and Target are turning to India for answers is not a good thing. Almost half of IBM's $96 billion in 2004 revenues came from such services (since the company sold its $10 billion personal-computer unit to China's Lenovo this spring, services account for well over half of revenues). And even though the services business is less lucrative than IBM's other lines of work--operating margins in services were 25% last year, vs. 31% for hardware (higher once you subtract PCs) and a Microsoftian 87% for software--only services promise to provide growth on the massive scale that IBM needs to make shareholders happy. Luckily for IBM, Wipro has only 100 retail consultants. So far.

Fast-growing, increasingly sophisticated, low-cost generic competition--known in the IT world by the ugly name of commoditization--is a bugaboo that can keep a CEO awake at night. If you're IBM CEO Sam Palmisano, it's an old nemesis. In the form of the PC, commoditization shifted the center of computerdom away from mainframes--a hugely lucrative but now stagnant IBM stronghold-- and confronted the company with fierce new rivals like Microsoft, Intel, and Dell. Nowadays commoditization threatens services too, and not just at telephone call centers--even consultants are being paid India's low wages. Meanwhile, IBM's costs remain those of a mature First World corporation. It has about 260,000 expensive employees in the U.S. and other developed countries (the other 60,000 are in lower-cost regions) and 164,000 pensioned retirees, all quarterbacked from a gleaming modern headquarters on 432 acres of pricey Westchester County real estate.

Though they don't put it quite this baldly, Palmisano and other top IBMers think they're well on the way to having the problem licked. Interviews with the CEO's lieutenants (Palmisano answered questions via e-mail) reveal a strategy that not only challenges upstarts like Wipro directly by taking the low-cost model right back at them but also includes a dimension that is so original and so bold that it will either reenergize Big Blue's profits or undermine its vaunted status as the biggest company in IT.

Simply put, IBM is gambling that it can win by giving away crown jewels--precious intellectual property in the form of software, patents, and ideas. Spread enough of those riches around, the theory goes, and the entire industry will grow faster, opening new frontiers. That, in turn, should create opportunities for IBM to sell high-value products and services that meet the new demand.

IBM's response to the India threat has been swift. In April, after first-quarter revenues from services came in unexpectedly weak and IBM disappointed Wall Street's earnings expectations, the company eliminated 14,500 jobs, mostly in Europe. (It was the biggest job cut in three years: The company shut its European headquarters and moved most of the surviving employees out to the field, to what Palmisano calls "client-facing positions.") Next, the New York Times reported having obtained an internal IBM memo saying the company would hire 14,000 people in India this year. IBM calls the figure "exaggerated," but newcomers in any number would add to an already surprisingly large roster in developing countries. Of the programmers who write custom code for IBM's services group, about half--26,000 or so--are in India, Brazil, or China. "Strategic low-cost geographies" is the IBM lingo for such places.

India already accounts for the largest number of IBMers outside the U.S. (it recently surpassed Japan). In 2004, Big Blue acquired India's Daksh eServices, whose 6,000 employees operate call centers for companies like and Citicorp. Goldman Sachs calculates that by the end of next year, IBM Services' headcount in India will top 52,000. That would be more than one-fourth of all its services personnel and about one-sixth of IBMers worldwide. It would put IBM in India on a par with Wipro, the largest local software company, and make it bigger than Infosys and Tata Consultancy Services.

Growth in the developing world is a natural part of implementing a "global delivery model" for services, says senior vice president Bob Moffat. In July, Palmisano reorganized services, naming the 49-year-old Moffat one of three executives who will run it jointly. While the other two will oversee the delivery of services to clients, Moffat's job is to find efficiencies. He spent the past three years taking billions of dollars in costs out of IBM's physical supply chain--the delivery of parts and goods to and from factories and on to the customer. His mission now is to cut the cost of delivering services, even high-value ones, by tightening the "services supply chain." That mostly means people--getting the right ones to the right place at the right time. He has to extract every last penny of value from IBM's 260,000 developed-country employees if they're going to stay on the payroll.

Palmisano has picked someone he knows and trusts--Moffat has worked closely with him at almost every step of his career--to oversee what is in effect a redesign of half the company.

The corporate Mr. Fixit rises daily around 4 A.M. to do e-mail and make calls around the world. He works till after 7 P.M. He has an obvious intensity, but simultaneously can be disarmingly casual and interested in whomever he's talking to. That serves him well in his endless dealings with IBM employees, many of whose jobs he is likely to change before he's through.

Moffat has launched an ambitious project called Professional Marketplace, a database of IBM's talent. While it contains profiles of only about 35,000 employees so far, Moffat plans for it to show the skills, location, and availability of about 250,000--all of IBM's services, software, sales, and distribution folks. It should help IBM ensure that it doesn't send an overqualified employee to a job that could be filled by someone who costs less. The issue is especially critical for IBM's 60,000 consultants, who bill by the hour. And since the system will include all those lower-cost employees in India, Brazil, and elsewhere, it should also help guarantee that if a job can be performed there at a lower price, it will be.

Palmisano, 54, will need every dollar he can get. In a FORTUNE cover story last year (see "Can IBM Get Great Again?" on, he said he intended to grow IBM's revenues at least 5% each year and earnings per share at least 10%. (That revenue growth number may not sound so ambitious, but consider: For a company IBM's size, it means adding the equivalent of a Wipro every five months.) Palmisano also wants IBM's return on invested capital (ROIC), which measures a company's efficiency at making money, to significantly exceed the average for the S&P 500.

His performance so far? When he succeeded Lou Gerstner in March 2002, IBM was beginning the second of two years in which revenues declined. Then, thanks in part to a global rebound in technology spending, results improved sharply: IBM exceeded Palmisano's growth targets in 2003 and 2004. He especially excelled at delivering ROIC, hitting 29% for 2004, vs. around 10% for the S&P.

This year is still a question mark. Sales and earnings rebounded after that terrible first quarter, and key analysts are now increasingly optimistic. Looking ahead, widely respected IBM watcher Toni Sacconaghi of Sanford C. Bernstein & Co. recently wrote that "2006 could be a solid year for IBM" and predicts that the stock will outperform the market. In mid-August, Steve Fortuna of Prudential Equity Group called IBM a "compelling buying opportunity." The consensus of analyst estimates currently projects revenues from ongoing operations to rise 4.8% and 5% in 2005 and 2006, respectively, and earnings per share to rise 10.2% and 10.3%. Yet while Palmisano is mostly making his targets, investors seem unconvinced. The stock, which closed recently at $83 a share, is still more than 15% below where it was the day Palmisano became boss. (In April it plunged to just under $72.) This despite the fact that in that same period, IBM purchased more than $20 billion of its own stock as part of a continuing series of buybacks.

But, man, does Palmisano sound confident. In an e-mail to FORTUNE, he makes a remarkable claim: that by adopting the strategy he calls simply "openness," IBM has tapped a major new "spur to innovation itself." The company lavishes some $5.7 billion a year on R&D. By sharing its discoveries wisely, Palmisano says, IBM will "make the pie bigger," and the entire industry will grow faster. That will help replace the profits lost to commoditization.

Collaborating with customers and even rivals to invent new technologies is a big part of this sharing plan, whose first fruits are already apparent. In hardware IBM has co-developed with Sony and Toshiba a breakthrough chip called the Cell, which could eventually help transform all of IBM's computers (see following story). In software, embracing Linux and other open-source technology has given IBM new platforms on which it is building almost all its high-growth applications.

The idea that giving things away makes the pie bigger for everybody is not IBM's invention, of course. Ever since the Internet started connecting the world--individuals to companies to nonprofits to governments--online collaboration has been on the rise. The most dramatic example: the nonprofit open-source software movement, with its legions of mostly unpaid contributors around the world. The trend is also evident in blogs and even in the collectively produced Wikipedia online encyclopedia (see "Money Makes the World Go Round--Or Does It?" on Says Palmisano: "This isn't theory for us. Collaborative innovation today is crucial to every aspect of our business. We've learned how to deliver value within this kind of business system and how to make money."

A walking, talking embodiment of that is Jim Stallings, Palmisano's vice president for intellectual property and standards--a job Palmisano invented when he tapped Stallings for it last September. Stallings, 49, used to manage IBM's work with the Linux community; now he's the guy who figures out what Big Blue should give away and what it should keep. A buttoned-down former Marine Corps captain and 14-year IBM veteran, he's not the most likely high-tech Santa. Yet he's almost hypnotically good at explaining how IBM's plan works. As new standards evolve, for instance, "we've learned that every time the industry agrees on a standard, a whole new wave of growth happens right afterward," Stallings says. "So the IBM company says, 'We're going to be very smart about where we embrace these. It's not going to be after we've lost 30 points of market share.' " Among the standards currently in the works, IBM has especially high hopes for the push to create an electronic health record (EHR) for the medical industry. Widespread adoption of such a standard would make patient data more measurable and comparable, which IBM expects could drive substantial sales of computer products and open new opportunities for services.

The giveaways to open-source software groups, customer groups, universities, and other IT companies are surprisingly extensive and diverse:

âóè IBM contributed management software to the Apache Geronimo project, a coalition of programmers that aims to create an open-source version of software businesses need to run their most demanding applications.

âóè Stallings orchestrated the creation of what IBM calls a "patent commons." In January the company announced it was giving away 500 software patents, which it valued at over $10 million. They now can be used free by anyone working on an open-source project.

âóè IBM gave a retail-industry group rights to patents for Internet access in stores. The group aims to make it easier for members to pull together data about customers as clerks are checking them out.

âóè The company is lavishing money and expertise on UC Berkeley, MIT, and other top engineering and business schools to institute an entirely new academic discipline it calls Services Sciences, Management, and Engineering. The program, which has cost IBM more than $10 million so far, echoes one it ran in the late 1950s that helped launch the field of computer science.

IBM says there's no way to attach a precise dollar value to its giveaways--but FORTUNE calculates they're worth at least $150 million a year. And while sharing is not yet a universal part of the culture, the company has come a long way from the arrogant, standoffish, monopolistic IBM of yore.

What's in it for IBM? Big Blue seldom gives away a technology unless it has intellectual property and expertise that will enable it to make money if the technology is widely adopted. When IBM hands out tools to retailers, it often sells them additional software and consulting services. What's more, freebies themselves can be a potent weapon. Proprietary application-related software from Microsoft can be costly for companies; Geronimo is free. Take that, Bill Gates! Geronimo is the latest salvo in IBM's open-source challenge to Microsoft. It has also abetted the Linux operating system, a formidable rival to Windows, and Firefox, a popular challenger to the Explorer browser. (Palmisano is so proud of IBM's up-with-Linux strategy that the sole photograph adorning his personal conference room shows him and Linux creator Linus Torvalds.)

Paul Horn, who heads IBM's famed research labs, loves explaining the bang for the buck the company gets by employing 600 programmers who spend all their time improving Linux. Developing and maintaining a world-class operating system costs a minimum of $500 million a year, Horn says. IBM's 600 programmers cost far less than that, but total spending by all the companies and organizations that contribute to Linux is way more than $500 million. "So by working in an open ecosystem on Linux," he says, "we get a world-class operating system for our customers' needs at a small fraction of what it would cost us to do it on our own." IBM often funnels the dollars it saves into developing proprietary software that works on Linux systems. "That's a business model that's very hard to compete with," Horn says slyly.

IBM has also found giveaways to be a potent door opener abroad, where the company derives 63% of its revenues. Stallings has been to China four times this year and plans to go again in September, twice. He's working to convince policymakers and business leaders that using open-source software makes more sense than buying Microsoft's. This IBMer bearing gifts has found a receptive audience. IBM is helping build a giant network based largely on Linux to connect all of China's libraries. "They're saying, 'We're going to build this for the next 1,000 years, on these open platforms,' " Stallings says.

Horn and others are quick to point out that openness is a great strategy only so long as you keep inventing products that let you take advantage of the doors you've opened. Already IBM has racked up some conspicuous wins. Its strong position in Linux helped boost sales when its hardware wizards developed "blades," miniaturized Intel-based servers that can be combined side by side on a rack, saving space and energy. IBM's blades are especially Linux-friendly--they incorporate features, for instance, that enable Linux users to expand their blade installations quickly and easily. And while IBM's market share in ordinary Intel servers is only about 20%, it dominates the fast-growing $2-billion-a-year blade business, where it has about 45% market share, according to the IDC research firm.

That all sounds good and smart, but the outcome is far from certain. IBM is in a brutally competitive marketplace with a long list of rivals going after the same customer dollar. Dell, the king of the commodity PC, is a $50-billion-a-year juggernaut that at its current growth rate (nearly 20% a year) would surpass IBM in size by 2010. Lately Dell has been hungrily eyeing the services marketplace, where it has identified low-end work like computer maintenance--once a reliable redoubt of profit for IBM--as a major source of growth. Ask founder and chairman Michael Dell about IBM's ambitions to get most of its growth in services, and he is quick with a putdown: "It's hard to say you're going to help your customers reduce their own costs when you're the high-cost provider yourself." Meanwhile, at the high end of the services business, respected analyst Christine Ferrusi Ross at Forrester says her research with customers suggests that Accenture is seen as a better problem solver than IBM. At $14 billion a year, Accenture is less than one-third the size of IBM's services business but is growing nearly twice as fast.

That is not a forgiving environment in which to play Santa Claus. Even if IBM succeeds in spurring industry growth, there's no guarantee its shareholders will benefit--it risks ending up like Xerox, which saw itself pushed to the sidelines in the 1980s and 1990s even as the computer mouse, graphical user interface, and Ethernet--all brainchildren of its Palo Alto Research Center--galvanized the world of desktop computing. Xerox helped other companies get rich.

Palmisano argues that IBM will be able to create products and services that capitalize on new markets as they emerge. But even his own executives need periodic convincing. Two years ago, Horn notes, some insiders fretted that IBM might be "shooting itself in the foot" by giving its ideas away. But in the end, company researchers concluded that as long as IT remains hard to use, expensive, and labor-intensive, with customers continuing to need help solving business problems--and any computer user would guess that means forever--IBM will have the opportunity to thrive.

Perhaps the biggest challenge Palmisano now faces is inspiring his rank and file. After all, commoditization threatens their livelihoods, as the recent mass cuts in Europe made painfully clear. And IBM's strategy aims to make the industry change even faster--which could, perversely, ratchet up the pressure on its people. (It can't be reassuring to know that Bob Moffat's talent database can pinpoint an IBMer on the other side of the globe who is as smart and well trained as you--and works for less than half your salary.)

Despite employee unease and its limp stock, the signs look good for IBM. It has a new, high-profit mainframe and a new superpowerful chip that has already helped turn around its semiconductor business. When researchers ask customers what they think of IBM, the results have been getting better. Forrester's surveys of IBM customers show that satisfaction levels jumped significantly this year compared with 2004. (The only large provider with a higher satisfaction level was Infosys.) When Goldman Sachs asked large-company IT managers which outfits were gaining or losing their business in July, IBM was the top gainer in both services and hardware. Edging out Dell, IBM led the list in hardware for the first time since Goldman began conducting the surveys in 2001. That's certainly a start. All Santa needs to do now is bring investors some happy surprises.