BOOK EXCERPT
Engines of Tomorrow
How the World's Best Companies Are Using Their Research Labs to Win the Future
By Robert Buderi
Simon & Schuster
(C)2000 Robert Buderi
All rights reserved.
ISBN: 0684839008
Engines of Tomorrow: How the World's Best Companies Are Using Their Research Labs to Win the Future, by former Business Week technology editor Robert Buderi, examines the role of corporate research in some of the world's top information technology companies. Following is the first, overview chapter. The rest of the book looks at the evolution of the field, with an eye to explaining the "research bloodbath" of the late 1980s and early 1990s, and details the current management and practice of research at nine leading firms: IBM, Siemens, NEC, Lucent, General Electric, Xerox, Hewlett-Packard, Intel, and Microsoft.
CHAPTER ONE
A Matter of Death and Life
"Research is a high-hat word that scares a lot of people.
It needn't. It is rather simple... Essentially, it is nothing but a state of mind
--a friendly, welcoming attitude toward change..."
-- Charles Kettering
The resurrection of Bell Labs research was conceived on a warm August night, in a flooding of Arno Penzias's soul almost as swift and total as an epiphany. The year was 1989. The lab's Nobel Prize-winning executive was vacationing in Tanglewood, home to a series of summer music concerts in the Berkshires town of Lenox, Massachusetts. But he wasn't relaxing. A few months earlier, AT&T president Robert Allen had outlined plans for breaking the company into a series of business units designed to help it compete better in fast-moving markets--and Penzias, then vice president in charge of research, had spent much of the day on the phone, preparing for his upcoming annual budget presentation and fretting over how to keep his scientists comfortable while also proving to senior management research would pull its weight under the new structure. That evening, Penzias tried to relax by attending a Boston Symphony Orchestra performance. Only instead of enjoying the music, he couldn't shake the feeling that something had gone terribly wrong with his great research organization.
"My gut was starting to grab," Penzias remembers. "I was listening to Beethoven's Third and my stomach was churning and I didn't know what to do." Suddenly, it hit him: the core research side of Bell Labs--about a tenth of the overall organization--had in many ways become outdated. Since its 1925 founding, the great enterprise had enjoyed a long and glorious ride highlighted by thousands of patents and seven Nobel Prize winners, including himself. But AT&T was no longer a regulated monopoly, able to run research at a leisurely, academic-like pace--and instead of worrying about writing scientific papers or building the world's smallest laser, it needed to focus far better on business objectives.
Penzias had resisted accepting this reality for several years. But everything changed that summer evening, in light of Allen's recently announced decentralization plans and surrounded by Beethoven's powerful music. "I just couldn't do it anymore," Penzias relates. "Somewhere along the line the business units were going to need help, and we were going to have to do something about it...And I realized I had more power to make that happen than I had thought."
Within a few weeks, Penzias had begun plotting a dramatic overhaul of AT&T's research activities. It took months of careful and exhaustive study. Seeking to identify redundant programs and inefficient practices, he launched a massive review of all activities--calling in a management consultant for strategic guidance. Then, on July 20, 1990, he presented his makeover in Bell Lab's big auditorium, stunning those assembled by vowing to "reverse the trend toward university-like research" that had come to characterize the place. Penzias declared that he was eliminating entire laboratories and that the organization would shift many resources to software studies--curtailing much of the physical sciences explorations that had made Bell Labs famous. "AT&T's senior management continues to invest more than one million dollars every working day in our work, and they give us the freedom to use it as we think best," he explained to his researchers. "That gives me a deep sense of responsibility to use our resources and our freedom wisely."
The changes put in effect that day shook Bell Labs to its soul. Soon, Penzias's policies were attracting headlines, inciting cries of outrage from national policy leaders that the legendary enterprise was abandoning the kind of fundamental scientific investigations that had given rise to the transistor and other great advances.
Even as the Bell Labs research head carried out his plans, a similar story was unfolding at IBM's Thomas J. Watson Research Center in Yorktown Heights, New York. There, at the helm of another of the world's great corporate research organizations, James McGroddy was spearheading his own brutal self-examination with that same bad feeling in his gut: a growing awareness that despite its own sparkling history of five Nobel Prize winners, research was not making much difference to the company.
Ever since 1990, much like his counterpart Penzias, McGroddy had been working to couple activities more tightly to IBM's various businesses‹-concentrating more on providing customers with "solutions" rather than just technology. By late 1992, the coming year's budget under fire, he had also begun to streamline support services, eliminate redundant or seemingly dead-end programs, and find outside sources of funding by bringing in government contracts or launching spinoff companies.
Prophetically, with the company in dire straits, McGroddy felt certain chairman John Akers would soon be compelled to resign. Since any replacement likely would know little of the Research operation, he prepared a complete documentation of the value the division created for and provided to IBM.
The 16-page document was only a few months old when new chairman and chief executive officer Louis Gerstner, Jr. arrived on the scene. The former head of American Express and RJR Nabisco liked to move fast. On April 1, 1993, the very day his appointment was announced, he called senior executives, including McGroddy, to the big board room at IBM's Armonk headquarters. The new boss asked his execs for a ten- or twelve-page description of their segment of the company to bring him up to speed on Big Blue's place in the world--a first step in determining the hard measures needed to put the lethargic giant back on the fast track. In particular, he wanted to know the details of each business, the economic model it followed, its customers and competition, strengths and weaknesses. The reports were to be submitted within a few weeks‹and after Gerstner had time to study the documents, he would schedule a meeting with each executive to discuss things in more depth.
In essence, Gerstner sought the very kind of report McGroddy had recently wrapped up on his own--a well-written statement of organizational practices and goals, coupled with a frank assessment of strengths and weaknesses, and an action plan for the future. McGroddy went back to the Watson center and slashed four pages from their report--turning the document over to Gerstner a few days later. Research was handily the first IBM group to complete its assignment--and the Watson lab was Gerstner's first IBM visit outside headquarters. The chairman liked what he saw. At the end of the five-hour visit, he told McGroddy that he had long served on the AT&T board, where he had witnessed the troubles associated with a great research lab that made little impact on customers or the bottom line. "I want to thank you for not putting that problem on my plate here at IBM."
Over the next two years, driven largely by McGroddy's vision, Gerstner eliminated third of IBM's total R&D budget, moving from about $5.1 billion a year to a tad under $3.4 billion. It didn't matter that inside IBM the battle had been won--and that the new chief executive began featuring research as a key to the company's revitalization. Just as with AT&T before it, many leading academics and policy wonks concluded Big Blue was abandoning science, and with it the future. You could hear the screams from Harvard to Washington, D.C.
Jump forward almost a decade, to the last years of the 20th Century. Arno Penzias and Jim McGroddy are both retired, though active in a multitude of research and development issues. Picture Penzias, still a Bell Labs consultant, just arrived at lab headquarters in Murray Hill, New Jersey, from his home in San Francisco. The cathedral-like lobby harbors a first-hand compendium of the electronics age. Display cases contain the original transistor and a model of Alexander Graham Bell's pioneering telephone. A Telstar 1 satellite hangs from the ceiling, while a series of plaques describe the pioneering work of the lab's eleven Nobel Laureates--including Penzias's 1965 discovery, with colleague Robert Wilson, of the cosmic background radiation presumably left over from the Big Bang.
But the former research vice president pays them no heed. Still fit in his late 60s from swimming and running, a laptop computer slung over one shoulder, he strides purposefully past the guard counter into the facility itself--where it's a whole new day. Whereas many once considered Bell Labs a national asset: today's operation belongs solidly to its stockholders. Fundamental investigations have been scaled back, especially in Penzias's beloved physics. Arun Netravali, his chosen successor, emphasizes connection to business goals and customers. Projects lean to the shorter term, with many managers attuned far more to products than scientific accomplishments. And it's working. Under the tutelage of rising star Lucent Technologies--one of three companies formed by AT&T's 1996 self-inflicted breakup--the lab has unleashed a cavalcade of innovations that Penzias asserts "by my count have added tens of billions of dollars to Lucent." He gets good feelings from Beethoven these days.
Jim McGroddy seems just as content. While serving on a variety of small company boards and government or professional committees, he devotes much of his time to a non-profit business--Advanced Networking and Services--that advances high school science education. At the research chief's retirement party in late 1996, chairman Lou Gerstner told how he had considered breaking up the Research division and parceling its resources and personnel out to individual business units so that those assets could be brought directly to bear on Big Blue's product needs. But McGroddy's tremendous first impression convinced him to stay the course. Within a few years, the research ship had turned around almost completely. Three new labs opened. Meanwhile, despite what was widely perceived as savaging its R&D budget in 1993 and 1994, IBM began winning more U.S. patents than any group in the world--a streak that by 1999 had reached its sixth consecutive year. "People bitch about input measures," states McGroddy. "But where's the missing output?"
***
Corporate research is dead. Long live corporate research. From the hallowed, warren-like corridors of Bell Labs to the graceful stone-and-glass crescent of the Thomas J. Watson Research Center, from Xerox's famed Palo Alto Research Center cascading down the sunny California hills to the glimmering high-tech sheen of NEC's basic research lab in Tsukuba or the fish-stocked ponds of Siemens' sprawling Erlangen facility, come stories of trauma and renewal, death and rebirth.
The dark days lingered for years--as the research bloodbath in many ways spurred by the big early 1990s upheavals at Bell Labs and IBM rolled around the world. In 1992, after decades of growth that generally far outshot inflation, industrial research spending dropped in real terms. It continued to spiral down the next two years, an unprecedented period of decline that affected nearly every major research lab in computers, telecommunications, and...name-your-industry.
Those sweeping cutbacks have drawn impassioned cries of lament--and outrage. The biggest concern is nothing less than the fate of national economies. With research costs soaring, and global competition intensifying, various academics and policy pundits have warned repeatedly that corporations are putting the brakes on science. The gravest danger, the argument goes, is that companies have focused so much on the "D" side of R&D that they are forsaking the more fundamental "R" work that creates the breakthroughs needed to spawn new industries. The New York Times summed up these sentiments with a front page article late in 1996: "Basic Research is Losing Out as Companies Stress Results." The Times warned the resulting shortfall of new technology could one day "shackle the economy."
But the various forces at work have been widely misunderstood. The axed laboratories and slashed budgets that characterized much of the 1990s obscured a vitally needed realignment--and the revival of corporate research has been almost universally missed. Rid of many of their vestigial ways and bad habits, the best labs today have moved research into another dimension--shifting their orientation beyond the old standards of merely inventing things to the more ambitious problem of innovation, artfully described by Xerox PARC director John Seely Brown as "invention implemented."
This drive for innovation takes a different face at every company and evolved at varying times and rates for each. But everywhere the aim is to vanquish the old linear model that says ideas progress from research to development to manufacturing to market in favor of a much more dynamic enterprise that includes constant interactions along this entire chain. Innovation requires researchers to simultaneously seize responsibility for escorting creations through corporate and marketplace barriers, listen to advice from inside and outside the company, retool their work based on that feedback, and do anything else necessary to get products to the customer.
It's true that in the face of new realities, including higher costs and stiffer competition, company research arms have had to scale back some longer-range projects. A hard line was especially needed in the ferociously competitive and tumultuous computer, telecommunications, and electronics industries, where technologies have converged at lightspeed to bring about what economist Raymond Vernon calls the "spectacular shrinkage of space."
But while the pressures of rapid change have forced research to be far more tightly coupled to the here and now--or at least the sooner rather than the later--the change is actually positive. In Europe, Japan, and around the United States, companies are evaluating projects with greater care, finding more effective ways of conducting research, and bringing innovations to market faster, spurring economic growth in the process. Tom Anthony, who with more than 160 patents ranks as the third-most-prolific inventor in General Electric's glorious research history, puts it this way: "In the old era, if you did something the chance of a business using it was zilch, so you couldn't get that satisfaction. Now, if something works, there's a really good chance it will be used."
Even more, the best labs today have regained their equilibrium and begun to see potential opportunity and wealth in all the mayhem. As colleagues complain about brutal and often-unfair competition from the far corners of the globe, these leaders point to untapped resources, collaborators, and ultimately new markets. And while pundits scream about a short-sighted focus on incremental improvements to existing products that short-changes the kind of path-breaking investigations that gave the world the transistor, they stress the payoffs from making hard choices that fit the times.
But basic research--better terms in the case of corporations are "pioneering" and "strategic"--is far from dead. In fact, as companies recover from the financially brutal early and mid-1990s--industrial research spending in the United States has actually ticked back up almost five percent annually since 1996--one can even find clear signs of a resurgence in more fundamental pursuits. After big cuts early in the decade, IBM is once again cautiously increasing its long-term research. Hewlett-Packard has nearly tripled research spending in recent years while ramping up basic or pioneering investigations into atomic structure and chaotic systems. Microsoft plans to boldly expand its fledgling research organization to 600 staffers by mid-2000 as it explores such far-out software and computer science issues as advanced interactivity and decision systems. Meanwhile, in 1995 Intel launched a research arm devoted to long-term studies in such areas as computer architectures and user interfaces.
What's different today is that such far out prospecting exists in an era of increased attention to corporate relevance in which projects are chosen more carefully to bear on areas likely to benefit the firm. Xerox's John Seely Brown likes to talk of a "bold but grounded" approach to research. "You get steeped in the real problems of a corporation and then go to the root and reframe when necessary," he explains. To pull off such a feat, researchers must combine fundamental studies with routine interaction with customers and counterparts from all around the company in order to develop products that not only work--but are needed in the marketplace. The way things used to be, Brown notes, "We were these elite scientists sitting in this building inventing the future. Already, talking about 'inventing the future' smacks exactly of the ontological problem. We don't invent the future. We can help enact the future--but we must work with others in making that happen."
And what a future is in the works. The world unfolding today is at once fully wired--and wireless. It's an era of super-smart cars that anticipate when a driver's about to turn or change lanes and make sure it's safe--and of "calm computing," when network computers and web servers will be quietly and invisibly embedded in appliances and walls, allowing clocks to reset themselves after a power failure, or paint to detect intruders. It's a world of atomic scale transistors, spurred to life by that universal high-tech mantra: smaller, better, faster, cheaper. But it's also a time where the transistor has taken another path: bigger, worse, slower--only way cheaper. Under this vision, low-tech plastic transistors that can be mass printed without clean rooms, etching, lithography, and all the rest are used to smarten up new generations of toys, luggage tags, and appliances--or even teamed with sensors to flash up-to-the-minute freshness dates for drinks and vitamins that take into account storage conditions.
Above all, the world being created today is friendlier. Computers turn on as users sit down, talk to their owners, even recognize basic moods and do things like hold calls after detecting a look of annoyance when the phone rings. It's a world of easier scrolling and easier searching through the Internet, based on intuitive combinations of words and pictures. Web surfers have been freed from the computer terminal by systems that enable people to listen to the Internet and check e-mail over their cell phones or car radios. Then, too, there's the vision of the Personal Area Network: the basic idea is that the body's fluidity allows it to serve as a wet wire--enabling people toting a specialized smart card to transmit a digital aura that securely conveys their bank account, driver's license, access codes and so forth without the need to swipe cards through "readers." Imagine: you walk onto a train or plane and are billed automatically, or rent cars and check into hotels without ever stopping in a line, slowing down only to scan a computer screen to find your parking space or room number.
By all indications that's just a start of the fun in store, for change is coming fast--and, like some benevolent rust, good research today never sleeps. Those at the best labs not only accept the fact, they like it that way. At General Electric's research center overlooking the Mohawk River outside Schenectady, materials scientist Minyoung Lee says tough competitors drive him to greater accomplishments. "It's no fun when you compete with a dummy," the Korea native asserts. "Then you got no work to do."
The days of one-shift isolation, where scientists ponder problems in a university-like lab for eight hours, then place everything on the shelf until the next morning without worrying about competitors, are fast dying out. The new era is moving closer in select areas like programming to three-shift innovation. Computer whizzes attack a problem all-day long, hand it off to counterparts a few time zones away as their time in the lab begins, only to see it relayed eight hours later to a third shift in yet another part of the globe before returning to the roost to begin the next morning. "You have the program developed as the Earth turns," marvels Bell Labs executive vice president of research Arun Netravali, "and that's a very exciting concept." Exalts Robert Spinrad, vice president of technology strategy for Xerox,"I visualize an image of the globe turning in space, and we begin at the interface between morning and day as people dash off to work on the same problem. It's fabulous what's going on."
***
If this portrait of surging vitality and renewed spirit flies in the face of headlines bemoaning the effects of all the early and mid-1990s funding cutbacks and reconfigurations kicked off by research leaders like Arno Penzias and Jim McGroddy, it's because too many people don't understand the evolution that has taken place and continue to mourn the old regime.
Corporate labs evolved not to produce scientific breakthroughs, but to bridge the gap between science and technology and create a useful product. The first industrial research houses arose in the German dye industry of the 1870s, when manufacturers grew wary of their dependence on buying patent rights from independent chemists and realized the advantages of establishing their own laboratories. The practice soon spread beyond Europe. In late 1900, with Thomas Edison no longer a presence in company affairs and his key patents expiring, General Electric decided that its own lab could be critical to maintaining leadership in the electric lighting arena.
On the heels of the GE move, DuPont, Eastman Kodak, and others opened research arms. World War I saw a host of firms going full-out to fulfill military contracts and dramatically increased the scope of corporate research--largely to reduce U.S. dependence on imported dyes, chemicals, drugs, metals, and other products. The push to better control the future through research continued during the interwar years. New Year's Day 1925 saw the creation of the Bell Telephone Laboratories on the western border of Greenwich Village. By then more than 500 American corporations ran their own research shops.
Occasionally these labs engendered fundamental scientific breakthroughs. In 1932, for instance, GE's Irving Langmuir won the Nobel Prize for his contributions to surface state chemistry. But science was never the real aim of industrial labs. Rather, writes historian of technology George Basalla in The Evolution of Technology, the point was to employ "research scientists to advance industrial goals."
World War II, though, fueled another era in research. The success of the atomic bomb effort at Los Alamos and the even bigger project to develop microwave radar at the M.I.T. Radiation Laboratory set the tone for an upbeat style of research: fast-paced and involving large interdisciplinary teams of engineers, biologists, mathematicians, and physicists--both experimentalists and theorists. The war was all about applying their skills to turn out new and better weapons and technologies--and in the process scientists became national heroes, holders of the keys to the future.
As the Cold War heated up, warming with the first Soviet nuclear explosion in August 1949 and then sizzling with the 1957 Sputnik launch, the United States witnessed an unprecedented federal openness toward funding basic science in universities, government labs, and industry, especially when it came to farther-out projects with possible military applications.
Such factors helped spur phenomenal growth in the numbers of industrial labs--and with the great expansion came a far more science-friendly attitude toward corporate research. When General Motors and Ford opened major labs in the 1950s, they saw fit to probe subjects in basic chemistry and physics far beyond anything likely to impact automotive technology. In both cases, President Dwight D. Eisenhower spoke at the dedication ceremony via a special closed-circuit television broadcast, depicting the centers as nothing less than defenders of democracy. Pegged to the 1956 opening of the $125 million GM center, the New York Times Magazine ran an article entitled, Key Men of Business--Scientists. "The welcome mat is out for eggheads in industry," it began. "With atomic energy, jet engines, automation and other scientific revolutions erasing the boundary between university and factory, scientists are becoming as fundamental as salesmen in industry's scheme of things."
As part of this new Ivory Tower climate, lounges were stocked with tea and cookies to encourage interaction between staffers. Especially at the biggest facilities, scientists found a unique freedom to pursue ideas. Researchers basked in this freedom, often tossing inventions over the wall to development without regard to making them work in the real world.
The era of science worship had reached full swing when AT&T scientists Arno Penzias and Robert Wilson measured the cosmic background radiation in 1965, and IBM physicist Leo Esaki and GE's Ivar Giaever made pioneering investigations in so-called tunneling phenomena, for which they shared the 1973 Nobel Prize in physics with Brian Josephson of Cambridge University in England. The idolatry rolled on as far as the late 1980s, when four members of IBM's Zurich Research Laboratory won Nobel Prizes: Gerd Binnig and Heinrich Rohrer in 1986 for their invention of the scanning tunneling microscope, followed the next year by Georg Bednorz and K. Alex Müller for their discovery of new high-temperature superconducting materials.
But corporations need far more than good science to thrive in the marketplace--and already by probably the mid-1960s a few companies were beginning to realize that something was wrong with the innovation machine. Over the next two decades, as U.S. firms found themselves clearly outmaneuvered and out-innovated by a host of Japanese companies in everything from cars to televisions to Walkmans. the message came through loud and clear.
Much of the attention focused--and rightly--on the research and development pipeline. The very hugeness and scale of R&D in the United States, and a widespread lack of communication between a company's research, development, manufacturing, and marketing arms, had created vast cultures of inefficiency. Even when researchers weren't busy worrying about their scientific careers and concentrated on creating new or better technologies, the often chasm-like divides between the labs and the corporate mainstream made it extremely difficult to consistently translate good research into products.
By the late 1980s, with many companies moribund and widespread layoffs the rule, corporate research seemed poised for outright disaster. The nineties opened with AT&T slashing R&D spending 8 percent--before adjusting for inflation. Aerospace giants Lockheed, General Dynamics, and Northrop also showed double-digit dips in their research and development outlays. The next year saw big cuts at General Electric. Then in 1992, embodied by Jim McGroddy's axe-wielding over at IBM, came the full-borne bloodbath that extended to companies around the world, and in nearly every industry.
In all these venues the particulars may have been different but the central aim was the same: bring down barriers separating research and development, speed up technology cycle times, and work more closely with business units to hasten products to market. These same essential goals continue to dominate research agendas charging into the millennium. That's because in these heady times of cell phones and wireless Internet links, Dells and Gateways, virtually all the factors that contributed to the rise of industrial Ivory Towers have evaporated. Consider:
-- Corporate research blossomed in companies like GE, IBM, AT&T, and Kodak, whose dominant market positions or outright monopolies cushioned budgets from narrow margins and allowed ample room and time to fully profit from new technologies. Today, with deregulation and global competition the norm, companies essentially have no alternative but to hold research operations much more accountable. As Arno Penzias puts it: "The genie of international competition has escaped from its bottle and shows no signs of returning...With billion-dollar contracts riding on who can supply the best technology at the lowest price, today's scientists can hardly afford to behave as they did when we were the only game in town."
-- A dramatic increase in the volume of scientific and technical research also makes it difficult to get a clear edge from fundamental inquiries. Andrew M. Odlyzko, head of AT&T Laboratories's mathematics and cryptography research department, notes that when IBM researchers Bednorz and Müller announced high-temperature superconductivity in 1987, rival groups at the University of Alabama, the University of Houston, and Bell Labs pushed the field forward with further discoveries in only a few weeks. In the past, research planners could count on a reasonable shot of turning a scientific breakthrough into a dominant market position. However, Odlyzko asserts, "These assumptions are no longer believed by industrial R&D managers, and are being questioned by national policy makers."
-- Many firms these days succeed despite doing little research.In the case of personal computer makers, the top three market leaders in 1998--Compaq, Dell, and Gateway--have comparatively little technology of their own. Instead of research, their innovations lie in more business-oriented realms such as manufacturing, marketing, and distribution--adding to the pressure to compete in real time and not far down the road. As erstwhile Xerox research head and Sematech chief executive William Spencer has joined Harvard Business School professor (now professor emeritus) Richard Rosenbloom in noting: "The success of these free riders deters further investment by traditional pioneers in industrial research."
Coupled with rising costs of doing science and the end of steam-rolling growth in federal support of research (in the United States, at least), these pressures have spurred the need for a more applications-oriented strategy--a reality that is here for the foreseeable future. Although industrial scientists may want to be valued solely for their knowledge, Lucent Technologies theoretical physicist J. C. Phillips told a European Physical Society meeting, "Specific industries...cannot achieve prosperity in the highly competitive markets of an overpopulated world simply by consulting physicists who understand general principles. The physicists must be eager to take their understanding and use it to develop specific products."
In light of such realities, research managers have had little choice but to question basic assumptions about how they went about their jobs--even what constitutes good research. One of first things to go as budgets tightened was the old notion that more is better--both in terms of overall research expenditures and the number of projects supported. That meant focusing activities on areas of strategic importance, often times core technologies in which the firm excelled and that applied to a wide range of products or business lines. It has also dictated making research far more accountable to the bottom line. "Basically, we are not a profit center, we are a cost center," notes Thomas Grandke, president of Siemens Corporate Research Inc. in Princeton, one of the German giant's three main research houses. "It requires a continual justification that we are worth our money."
It was primarily the need to justify the existence of Bell Labs's research that prompted Arno Penzias's Tanglewood epiphany. His right hand man, Bill Brinkman, was speaking specifically about Bell Labs--but he might just as well have been summing up the state of research at scores of other corporations when he said: "What was very clear ... was that we were in deep trouble. The business suits in AT&T were really coming at us and saying, 'Hey, what in the hell are you doing for us?' The answer was we were just too researchy."
***
Rather than marking the end of corporate research, the transformation that has rumbled through labs around the world heralds a new era--maybe the most exciting yet. The watchword for the millennium is "relevance," and what the historian of technology David Hounshell calls "a crazy quilt of avenues, approaches, and opportunities" has emerged to augment the innovation process and ensure that what's conjured up in labs matters to the parent company and its customers.
Research organizations today act more quickly to jumpstart promising projects--often affording them special fast-track status. But they're also faster to pull the plug if things seem to veer off target. At the same time, technology is so diversified and specialized that companies can no longer do everything in-house. So they band together to form consortiums and R&D joint partnerships, contract for work from private institutes or government labs, license technology, buy start-ups, bankroll start-ups of their own, and fund university studies. Tying everything together, teleconferences and virtual meetings via the Internet link scientists and engineers across nations and continents--and the day is coming fast when every computer will have its own video camera to allow this interaction to occur more easily and spontaneously.
Some of these measures or tools have been around for years, others are brand new--and still others represent mutations of old ideas. And while many steps are small and commonsensical, a growing number flip conventional views on their heads. Lee Davenport spent 15 years as research director for General Telephone and Electronics in the 1960s and 1970s, taking part in numerous government committees addressing R&D issues and establishing himself as a leading voice on corporate innovation. The best organizations, he maintains, cultivate innovation by creating a climate that connects staffers to the real world, but also encourages out-of-the-box thinking. "Research is never just a gamble," he says. "You can definitely shape the odds in your favor." In a series of lectures at the University of Virginia's Darden Graduate School of Business Administration, he spelled out seven basic steps to successful research that remain essential guidelines for corporate labs.
Shaping the Odds: Lee Davenport's Seven Rules for Innovation
It's not easy bringing good things to life. Finding standout researchers is only a first step. Far more difficult is turning exciting ideas into things that matter. A world of strategies tries to bring this about--from touchy-feely teamwork sessions to hard-wired number-crunching. These can work, or fail, depending on lab style, culture, and implementation. The best companies, though, seem to share a few simple measures that forge a framework for innovation.
Octogenarian Lee Davenport restores vintage cars and caravans in road rallies. But as a physicist and industrial research director he has enjoyed a ringside seat on the electronics age--from tubes to chips, analog to digital. In war and peace, he's seen ideas come and go--and come again. That gives him an all-too-rare commodity: perspective.
During World War II, at the M.I.T. Radiation Laboratory, Davenport helped create a revolutionary fire-control radar instrumental in shooting down buzz bombs over England. After the war, he spent 15 years as research director for General Telephone and Electronics, now GTE Corp.--a major local phone service provider that in those days owned Sylvania. Based on his lecture notes from the University of Virginia's Darden School of Business, he culled out seven common sense rules of corporate research proven to set the stage for innovation:
1). Success is based on schedules and results--not effort, job difficulty, or loyalty. "You must expect your R&D people to produce results and reward them accordingly."
2). Since most projects last several years, managers must break them into shorter segments, with measurable goals at each phase.
3). Never allow general goals. Avoid such words as: approve, advance, increase, investigate, study, explore. All are false goals--immeasurable.
4). Look for idea people. Only a few individuals have truly unique, even hare-brained ideas. Encourage them.
5). Find product champions--internal entrepreneurs who understand technology, explain it clearly, and can push ideas through corporate barriers. These traits typically elude top researchers.
6). Keep a little something on the side. A bootleg research budget is sometimes the only way to pursue ideas that break the mold.
7). Hire young blood. A research staff's average age must not increase even one year per annum. In a high-tech lab, a nice average is under 35.
|
The vigor and vim of corporate research stands out clearly on the broader global canvas. In recent years, competitors have dosidoed into each others' beckoning markets--greatly leveling the playing field in Asia, Europe, and the United States. At the same time, more equal science and technology education--combined with a growing tendency of first-rate scientists to stay in their home country--has distributed the talent pool far beyond American borders, where it had been largely contained in the first three post-war decades.
Top research-oriented companies have tried to tap these far-flung resources by opening labs around the world. Operating these outposts can add geographic, language, and cultural barriers to the innovation problem. American High Definition Television expert Jack Fuhrer was hired by Hitachi in 1991 as senior director of its new Digital Media Systems Laboratory in Princeton. It took Fuhrer two years to realize that his Japanese managers didn't like to say 'No' outright, but instead relied on subtle clues such as a lack of enthusiasm or silence to get the message across. One day, though, Fuhrer was stunned when a senior manager flatly declined a suggestion. "I said, 'Matsura, this is weird, you don't usually say no.' He said, 'I'm trying to help. I've learned when I think 'no' it helps you if I say 'no.''"
Despite such pitfalls, though, foreign labs are better attuned to the needs of local customers and markets than the mother ship back home--and can more easily leverage developments in nearby companies and universities. They also offer a welcome source of diversity by bringing into the innovation stream people with varying cultural views and perspectives that add rich dimensions to attacking problems. That's why the tendency to go abroad has picked up steam in recent years. Beyond home shore labs in New York, California, and Texas, IBM runs research operations in Israel, Japan, Switzerland, China, and India: the last two have opened since 1995. Underscoring Microsoft's newfound commitment to research was its 1997 decision to invest $80 million in a new facility in Cambridge, England, a move followed late the next year by the creation of a Beijing lab. "The market environment is different among the United States, Japan, and Europe--and each region has its own strengths, its own characteristics," notes Tatsuo Ishiguro, associate senior vice president in charge of research and development for NEC, which has opened labs on all three continents in the past decade. "So by mixing these different strings we want to get higher, or stronger."
But starting labs is just one facet of what's happening. More telling, perhaps, is where companies put their researchers. For some forty years beginning in the 1950s, the trend for major corporations was to place labs near academic institutions or in secluded, pastoral havens. But when Microsoft launched its central research effort in 1991, the facility went up on the company's main campus in Redmond, Washington, where it was easier to build ties to the product development groups next door. Intel chose a similar path when creating its far-looking Microcomputer Research Laboratory four years later. Other companies worry about researchers throwing inventions over the transom to developers. In Intel's case, there was no transom. Research set up shop on the sixth floor of company headquarters, with nary a sign to distinguish it from the developers and business groups all around.
The assault on time-honored practices extends to all levels of research--down to the workbench. A prevailing theme these days is to hand scientists more responsibility for getting their creations to market. Before licensing its technology to Norton, IBM launched major updates of its AntiVirus application directly from research. Theorists, physicists, programmers, and interface designers all worked together on the program, which includes a neural network to detect previously unknown viruses. In 1997, this ad hoc collaboration brought the latest version of the package from conception to market in five months, the second time in the product's five-year history that researchers had performed a similar feat. Jeffrey O. Kephart, manager of IBM's Anti-virus Science and Technology group, called the process "almost instant gratification. You can have a seemingly wild, improbable idea, and a few months later millions of people are benefiting from it."
Xerox PARC has added yet another twist to the idea of linking researchers with markets. As in most labs today, scientists and engineers routinely bring clients to the lab, and in turn visit businesses in order to better understand what companies want. But a long time ago PARC managers realized that while computer jockeys and other technologists are usually adept at seeing how their products and expertise can help fill customer needs, even more value can be created by also letting customer needs shape their inventions. Today, the center's 300-odd researchers include a half-dozen anthropologists who have gone into a law firm, airline operations room, and government engineering offices to study the hidden and intuitive ways people do their jobs--and then work with computer scientists to create novel technologies designed to facilitate those often-overlooked workstyles.
These strategies and initiatives mark only the tip of the innovation iceberg. To help spur creativity, Siemens sponsored a company-wide ideas contest. Through thick and thin, Hewlett-Packard has maintained its long-standing practice of allowing researchers to devote 10 percent of their time to unofficial, bootleg-type projects. Each year IBM grants some products or developments special status by deeming them an Accomplishment, Outstanding, or, at the highest level, Extraordinary. Ratings depend either on the innovation's general impact on science and technology, or its effect on IBM's business. The latest Extraordinary rating came in 1998, when the SP-1 scalable programmable parallel computer known as Deep Blue--the machine that demoralized world chess champion Garry Kasparov in 1997--was elevated in rank from Outstanding.
Wielding all the weapons in the contemporary research arsenal may not be relaxing--but it certainly adds excitement. What's especially different these days is the sheer number of these tools and strategies, far more than available to the previous generation of managers. Raising the stakes even higher, all must be interwoven and operated in parallel--for everything takes place in the context of an intensely competitive world where alliances, products, fashions, prices, and just about everything else change colors faster than a chameleon.
Heading into the 21st Century, Bell Labs research director Arun Netravali cites three fundamental tenets of successful corporate research: speed, complexity, and cannibalization. Largely due to competition, the need for speed is paramount--in evaluating projects, pursuing advances, adopting outside technologies, and creating novel products. Complexity is another way of saying that nothing is straightforward. For example, Lucent's business hinges on data switching and transmission. Now that PCs are often connected to local area networks, data packets can be transmitted at incredibly varying rates--from a snail-like one an hour to a blurry million per second. The seemingly incalculable degree of randomness--combined with the fact modern systems utilize equipment from a series of vendors, all of which must be maintained with incredible reliability--has created data management headaches orders of magnitude greater than just a few years ago.
Overcoming all that, though, is almost the easy part. The next step is to obsolete everything: cannibalization. This goes far beyond simply coming up with subsequent generations of products to eating your own lunch. For instance, modems have evolved from boards to chips. Yet at AT&T, before trivestiture, the organizations stamping out modems and chips were entirely separate. So in some senses chip-makers had to think about rendering their counterparts obsolete. Says Netravali, "Let us become better at doing this than some outside company, because it's going to happen anyway."
But is there a price to pay? The metamorphosis in corporate research aims to cement ties between scientists and the world outside the lab--an action everyone associated with the field calls warranted. The big worry in academic and policy circles, though, is that as companies move toward such product-centric work they will abandon the risky fundamental studies that often ignite future growth. In the 1980s and 1990s, for example, General Electric, Siemens, and a number of other major corporations retooled budget formulas so that research labs now depend on contracts with their own firm's business units for much more of their funding than in the past. The contracted work almost never involves anything more than a two or three years out.
"The fact is we are no longer getting the major contribution of long-term research from the corporate labs that we had done for decades," asserts Allan Bromley, science advisor to President George Bush. Of investigations such as those leading to the transistor, laser, and fiber optics, he states, "That has simply been cut out altogether." Former Sematech head Spencer, joined by Harvard's Rosenbloom, sees the recent trends as transforming corporate research centers from prime sources of innovation to the pipeline through which corporations filter innovations from the outside. "Without a concerted effort by government, industry, and universities, the engines of innovation that have worked so well for the United States during the past five decades could be significantly impaired," the two conclude. Going a step farther, Edward E. David Jr., another former presidential science adviser who also directed Exxon's research, once predicted that if the drive toward incrementalism and applied R&D were continued to its logical conclusion, central labs would be eliminated and research enslaved to technical service.
These views--or fears--are shared by many others. But while central labs certainly aren't for every company, the institution is not going away--and in some ways is becoming more important. In November 1998, out to strengthen its handle on the future, Motorola boosted research spending and combined previously separate investigations in semicondutors, wireless communcations, and other fields into a new organization called Motorola Labs. Innovative thinkers like Xerox PARC director John Seely Brown go so far as to assert that this type of centralized research holds the power to "reinvent the corporation" and provide the "genetic variance" that ensures a company's long-term survival.
It's not that big companies expect their central labs to be first to market with new products: that rarely happens. Instead, argue research managers, the purpose of places like Bell Labs and GE's historic Schenectady complex is to keep their parents in the game for the long haul, by providing a full range of integrated services and technologies that customers can count on with no significant loss in lead time. This is possible because central labs not only harbor a well of expertise to evaluate and assimilate the rising tide of outside inventions, they possess the ability to supplement those developments with a stream of innovations and discoveries of their own. Venture companies may be more nimble than large organizations because they can focus on a single project, notes NEC research head Ishiguro, "but the merit of the big organization is that we have a wide range of technologies. If we can get synergy of this wide range of technologies, we can have some merit or some strengths."
Xerox vice president of research and technology Mark Myers, who oversees all the company's research and advanced technology operations worldwide, points out that central labs have also emerged as critical players in research alliances or consortia. That's partly due to their being well-situated to absorb the fruits of external research. But it's also because firms entering into such increasingly important partnerships must typically bring something to the table: often that is proprietary technology developed in the central lab that is broadly applicable to the entire industry. Furthermore, says Myers, unlike research operations tied to individual business units, a central organization offers the unique capability to look across product lines, identify the critical forces of change, and shape the firm's strategy to meet the likely consequences. Finally, being a step removed from market pressures, its managers are often more ready, willing, and able to commit strategic resources to longer-term issues.
The extended time horizon of central labs is why many directors also insist that basic research is alive and well--if not thriving. The 1990s have seen corporate labs cut back significantly on fundamental studies. A first step came in reining in "pure" science--that is, pursuits characeristic of a university aimed solely at advancing the understanding of how nature works. In truth, only the biggest and most dominant labs, such as IBM and Bell Labs, ever engaged in this kind of basic research--and it was never more than perhaps one percent of the total research effort. Far more typical, but still only a tenth of the budget at major labs, were quasi-fundamental investigations such as the solid-state studies leading to the transistor. Although they can advance science, these "strategic" or "pioneering" efforts are a step removed from "pure" studies because they are well-focused on areas likely to benefit the company.
Many of these pioneering investigations have also been phased out. But those painting a dark picture of the future of fundamental research misread the situation. Theirs is an overly impatient view that does not allow time for firms to respond to the acute need to cut costs, speed up research and development cycle times, and come to a new equilibrium. John Armstrong, former IBM vice president for science and technology, likes to say the critics suffer from "binary syndrome." That is, they think basic or strategic research is either on or off--and that "if it isn't the same or greater, it's zero." Moreover, these critics fail to realize that by refining areas of focus, managers can often improve their odds of hitting a home run.
In today's research world, with the merits of different types of research--basic, strategic, pioneering, applied, long-term, short-term--being debated up, down, and sideways, and in several cases blurred almost beyond recognition, some favor doing away with these terms entirely. "I do not like this differentiation between basic research and applied and fundamental, et cetera," asserts Claus Weyrich, who as head of Corporate Technology for Siemens oversees its three central labs worldwide."I'd like only to differentiate between good research and bad research." Bad research helps nobody, Weyrich says. Good research is what helps the company, no matter its nature or time horizons.
That being said, even though sustaining vigorous programs of fundamental research is typically viewed as risky, another way of looking at the issue is that not pursuing such studies means rolling the dice in another way: by forsaking the future. That explains why the best companies still walk the fine line between short-term research and the need for farther-out, path-breaking investigations.
While successfully implementing Arno Penzias's moves to shed much of its academic style, Bell Labs has not cut the overall size of its Physical Research Laboratory, where some 140 scientists pursue investigations not likely to bear fruit for two decades or longer. Physical Research is where Alan Gelperin experiments with live slug brains to uncover basic principles for constructing biological computers, astrophysicist Tony Tyson maps the universe's invisible dark matter by studying the ways its gravitational forces bend light, and Federico Capasso pursues his invention of the quantum cascade laser, a novel and highly sensitive device with applications in everything from detecting illegal drugs to analyzing auto pollution.
It's the same at other top labs around the world. In the Japanese high-tech enclave of Tsukuba--dubbed Science City--NEC researcher Sumio Iijima strives to advance his discovery of carbon nanotubes, strings of super-strong carbon atoms whose unique electrical properties might be perfect for creating nanometer-size transistors. On the other side of the world, at IBM's famous research laboratory in the hills above Zurich, James Gimzewski plots to fabricate molecular machines too small to see with the human eye, showing off pictures of history's smallest computer--an abacus he fashioned out of buckyball molecules.
Indeed, some firms which never engaged in basic or strategic investigations are now taking the plunge--though with the 21st Century twist of asking researchers to think more about potential commercial payoffs than in the past, when corporate scientists were judged largely on the volume and importance of their published papers. Take Microsoft natural language processing expert Lucy Vanderwende, who pursues fundamental linguistics studies in an effort to enable computers to understand everyday human speech and respond in kind. Vanderwende and her colleagues have constructed a Product Rainbow, which consists of two arcs laying out milestones they hope to reach at various stages. The "rainbow's" inner arc describes the basic scientific goals such as morphology and logical form. Just outside it, another arc details some products the group expects to create as it achieves the scientific goals--including a grammar and style checker that debuted in late 1998.
Hewlett-Packard, longtime king of the applied labs, also has dramatically expanded such pioneering studies in recent years. The establishment of the Basic Research Institute in the Mathematical Sciences at its lab in Bristol, England, marks just one example. There, theoretical physicist Jeremy Gunawardena supervizes work that includes probing the possibilities of building quantum computers vastly more powerful than today's supercomputers. To achieve that goal, he relates, scientists will have to resolve some basic questions debated by Albert Einstein and Niels Bohr. "That's really when the physics changes. It changes from being semiconductor physics which people have known and studied for 40 years and really brings you up against some foundational questions in quantum physics."
That's just the kind of basic research Joel Birnbaum, H-P's senior vice president of R&D, might have encountered years ago when he worked for IBM. The difference today, he relates: "We don't say, 'Think great thoughts for the rest of your life.' We say, 'We are working on topics in early stages of fundamental understanding. We hope you'll be very disappointed if people don't apply what you've learned in 10 to 15 years.'"
***
Ralph Gomory, a mathematician who led IBM's esteemed research organization for 17 years during the heyday of science, tells the story of growing up in Brooklyn Heights in the 1930s. Neighborhood kids liked to congregate on a one-block-long street called Garden Place, where they played stoop ball, bunch ball, and roller-skate hockey. Parents never worried about their safety. No cars were parked on the street, and few even drove down it. Over time, though, traffic gradually increased. It was hard to notice each change as it happened, but now Garden Place is lined with cars, and somewhere along the way it became impossible for kids to play those same games. "It was unplanned evolutionary change," notes Gomory. "Yet the whole structure of life for children changed."
A lot of that same kind of unplanned evolution has taken place in corporate research. A few decades ago, Gomory notes, corporate labs needed to offer some researchers a great deal of freedom and latitude to pursue fundamental studies: otherwise, they wouldn't have come. But the world has changed, dictating pursuits more directly relevant to company aims and goals. Just like in his old neighborhood, Gomory relates, people have to face that reality. "There's no choice."
Maybe it all could have been done more peacefully, smoothly, elegantly, caringly, sparingly--a kinder, gentler end of an era. After all, notes Harvard Business School's Richard Rosenbloom, "The proper way to run a research organization is on a steady course." If labs undergo cycles, swings, and traumatic change, he asserts, "I think that's a confession of major error." And indeed, Rosenbloom warns that the true consequences of the shift and cutbacks in longer-term research might not be visible until early in the 21st Century.
There's a lot to be said for that point of view. Places like IBM and Bell Labs felt the pain deep in their souls, whereas Hewlett-Packard hardly experienced anything at all--so it's possible to argue that an H-P-like, steady-as-she-goes management style could have minimized the trauma at other labs. Still, some things are outside anyone's control--and the changes in corporate research have proven so universal that it's hard to simply blame management. Even if companies made serious mistakes--and many did--it doesn't alter the fact that corporations must adapt to face a rapidly evolving world. So while naysayers in government, academe, and even corporate labs themselves often lament those changes--especially where it relates to fundamental research--the same truths that reshaped Ralph Gomory's old neighborhood have worked their will on corporate research. It's time to stop looking back.
Even more than that, the need for reinvention has always been part of life in the research world--as it is in any dynamic enterprise. Or, put another way, every healthy organism invests periodically in self-renewal. Just as Charles Kettering warned many decades ago, the alternative is to be washed away in "advancing waves of other people's progress." Of all the household names in consumer electronics from the vacuum tube days, only Motorola remains as a leading maker of electronic components. It had to shed traditional businesses and reinvent itself several times, moving from mobile radios to portable radios to pagers and cell phones--all activities in which research was heavily engaged.
That's what the best are doing today--using their research operations as the linchpins for transformation. It's even possible that to the extent labs get more connected to business needs, and shed themselves of the old academic style while still maintaining some far-ranging studies, the chances of making path-breaking discoveries actually increase. The reason: science alone was never what made places like Bell Labs and General Electric truly great. The truly special aspect of these places was that more than almost any university, the labs brought together world class scientists with experts in such areas as electronics and antenna design, theoreticians with experimentalists, chemists with physicists and engineers--and from that mix rose a tremendous spark of discovery.
Bell Labs astrophysicist Tony Tyson says the dynamic for such a spark may actually be better now than at any time since the 1950s--at least where he works. An increased focus on relevance has put short-term pressures on researchers and made it harder to pursue "pure" science. However, he states, "I think it's healthy to have this tension. Otherwise you're just sitting in the Ivory Tower doing nothing for anybody. It really does help to be immersed in the needs of the corporation at the same time you're trying to make some new discovery. If you're immersed in other cross streams of technology, of ideas, of demands...that's a very rich environment for completely new ideas to spring forward."
At the dawn of the 21st century, the essence of the inherent vitality of industrial research is often lost in a confusing eddy of forces. A first step in making sense of the current times is to resurrect and examine the past with a contemporary eye. Indeed, when tracing the evolution of industrial research from its German roots to the present day, it becomes clear that in many ways corporate research is actually more connected to its original purpose of advancing company goals than at any time in the past half century.
| CONTENTS
|
| Introduction: Change |
| ONE: A Matter of Death and Life |
| TWO: The Invention of Invention |
| THREE: Houses of Magic |
| FOUR: Out of the Plush-lined Rut |
| FIVE: IBM: Taking the Asylum |
| SIX: House of Siemens |
| SEVEN: NEC: Balancing East and West |
| EIGHT: The Pioneers: General Electric and Bell Labs |
| NINE: Children of the Sixties: Xerox and Hewlett-Packard |
| TEN: The New Pioneers: Intel and Microsoft |
| Conclusion: The Innovation Marathon |
From ENGINES OF TOMORROW by Robert Buderi. Copyright © 2000 by Robert Buderi. Reprinted by permission of Simon & Schuster Inc.
For more information about Simon & Schuster, go to SimonSays.com
Click here to contact Robert Buderi or to visit his Web site for Engines of Tomorrow
|
=Subscribers Only
Copyright 2000-2002, by The McGraw-Hill Companies Inc. All rights reserved.
