Only the Paranoid Survive

by Andrew S. Grove

  Technology can help you here. E-mail gives us a very powerful new way to reach large numbers of people. In most modern organizations, every computer is connected to a corporate network through which it can send messages to every other computer on that network. By spending minutes in front of his computer, a manager can leave his mark on the thoughts of dozens, hundreds or even thousands of people in his organization and do so with the immediacy that only the bounce of electronic communications can generate.

  One caveat: if your message is clear, it will provoke questions and responses which will come back to you through the same medium. Answer them. You don’t have to make a lot of time for that; a response of two or three lines can communicate the essence of your position. This is a high-leverage activity; your message will reach not just the individual to whom you are sending it, but is likely to be sent on to other employees of the network, reverberating, as it were, from computer to computer. So consider this as the electronic equivalent of answering a question at an employee forum. Be crisp and to the point, and your response will help move the thinking of employees toward the desired direction.

  I spend some two hours a day reading and responding to messages that come to me from all over the world. I don’t generally read them all in one chunk but I almost always make sure that I’m caught up before my day is done. I have found that I can project my thoughts, reactions, biases and preferences very effectively through this means.

  Equally important is the opportunity that incoming e-mail gives me to be exposed to the thoughts, reactions, biases and preferences of large numbers of people. More Cassandras bring me news from the periphery this way than through any other means. I witness more arguments, I hear more business gossip, sometimes from people I have never met, than I ever did when I could walk the halls of the one building that housed all Intel employees. What used to be referred to as “managing by walking around” has to a large extent been supplanted by letting your fingers do the walking on your computer keyboard. Given that Intel has now spread all around the world and I couldn’t walk the halls of our sixty-odd buildings even if I spent full time at it, this has become doubly important.

  A lot of times, management attempts to communicate a new strategic direction by the medium of closed-circuit television or prerecorded videotape. This may seem like a logical and easy thing to do but it won’t work. The interactive element, the give-and-take, the opportunity to ask the “Does it mean that …” type of question is lost in this one-way medium. If your employees don’t have an opportunity to test your thinking in live sessions or electronically, your message will seem like so much hot air.

  Resist the temptation to do what’s easy here. Communicating strategic change in an interactive, exposed fashion is not easy. But it is absolutely necessary.

  Adjusting to the New

  Gordon Moore’s comment that half our staff would need to become software types in five years’ time was a very valid insight and one that has an equivalent in every company that’s struggling with a “10X” force. Simply put, you can’t change a company without changing its management. I’m not saying they have to pack up their desks and be replaced. I’m saying that they themselves, every one of them, needs to change to be more in tune with the mandates of the new environment. They may need to go back to school, they may need a new assignment, they may need to spend some years in a foreign post. They need to adapt. If they can’t or won’t, however, they will need to be replaced with others who are more in tune with the new world the company is heading to.

  In our case, we achieved the changes in our management along the lines Gordon suggested. To be sure, some managers left and were replaced by others from within Intel whose backgrounds were much closer to the new requirements. But most of us learned new tricks. For example, as I described, I invested a fair amount of time in learning about the strategies of the software companies in the personal computer industry and establishing relationships with their managers. Others undertook parallel transitions to new assignments. Some even took a step back to a lower-level assignment—a demotion—and, fortified by experiences from which they learned skills that were more appropriate to our new direction, later rose back again in the management ranks. We do this often enough that it is an accepted way for managers to learn the new skills they need as the company heads in a new direction.

  Intel isn’t alone in exhibiting adaptive behavior. A company that has consistently been able to adapt to new directions over fifty years is Hewlett-Packard. I had the opportunity to see how they go about it. In recent years the management of Hewlett-Packard decided to base their future microprocessor needs on Intel’s microprocessor technology. The implication of that was that their computer business, which now uses a microprocessor that they designed themselves, would have to grow to rely on a microprocessor that would be available not just to them but to their competition.

  This is a profound change for their business and it had to have been a very hard decision to reach. I witnessed some of the discussions in meetings we had with them, and catching even a glimpse of the process gave me a sense as to why Hewlett-Packard has such a spectacular record of navigating transformations. The discussions were rational, nonthreatening and slow but they steadily moved forward instead of meandering around in circles.

  Sometimes management may see the need for a dramatically different new direction but not be able to bring the rest of the company along with them. I have seen a videotape of John Sculley, CEO of Apple Computer from 1983 to 1993, telling a Harvard Business School gathering that two of the biggest mistakes of his career were not adapting Apple’s software to Intel’s microprocessors and not modifying Apple’s then revolutionary laser printer so that it would work with PCs other than those made by Apple. I was stunned when I heard this. Listening to Sculley gave me the impression that he understood the implications of the horizontal industry structure. It appeared that he just wasn’t strong enough to overpower Apple’s inertia of success that existed because of its fifteen-year history as a fully vertical computer company.

  Then there is the intriguing case of Wang Laboratories. Under the leadership of the founder, Dr. An Wang, the company had gone through a tremendous transformation, from being a producer of desktop calculators to becoming a pioneer in distributed word processing systems. Dr. Wang understood these technologies and had an extremely strong hold on his company. His vision was law and his vision was generally spot on. But by 1989, when the PC revolution became truly significant, Dr. Wang was very ill. Without his firm hand at the helm and without senior managers who could step in and define the company’s new identity in this time of change, the company lost its strategic direction. This time the company didn’t make it through the transformation and actually ended up in Chapter 11 bankruptcy.

  Why couldn’t Apple and Wang rein in chaos?

  It seems that companies that successfully navigate through strategic inflection points have a good dialectic between bottom-up and top-down actions. Bottom-up actions come from the ranks of middle managers, who by the nature of their jobs are exposed to the first whiffs of the winds of change, who are located at the periphery of the action where change is first perceived (remember, snow melts at the periphery) and who therefore catch on early. But, by the nature of their work, they can only affect things locally: The production planners can affect wafer allocation but they can hardly affect marketing strategy. Their actions must meet halfway the actions generated by senior management. While those managers are isolated from the winds of change, once they commit themselves to a new direction, they can affect the strategy of the entire organization.

  The best results seem to prevail when bottom-up and top-down actions are equally strong.

  We can display this point in the following two-by-two matrix:

  Dynamic Dialectic

  The best quadrant to be in is the top right one—strong top-down and strong bottom-up actions roughly balancing.

  Dynamic Dialectic

  If the actions ar
e dynamic, if top management is able to alternately let chaos reign and then rein in chaos, such a dialectic can be very productive. When top management lets go a little, the bottom-up actions will drive toward chaos by experimenting, by pursuing different product strategies, by generally pulling the company in a multiplicity of directions. After such creative chaos reigns and a direction becomes clear, it is up to senior management to rein in chaos. A pendulum-like swing between the two types of actions is the best way to work your way through a strategic transformation.

  This dynamic dialectic is a must. The wisdom to guide a company through the valley of death cannot as a practical matter reside solely in the heads of top management. If senior management is a product of the legacy of the company, its thinking is molded by the old rules. If they are from the outside, chances are they really don’t understand the evolving subtleties of the new direction. They must rely on middle management. Yet the burden of guidance also cannot rest solely on the judgment of middle management. They may have the detailed knowledge and the hands-on exposure but by necessity their experience is specialized and their outlook is local, not company-wide.

  I learned this lesson the hard way. Intel before the crisis of the mid-1980s had a strategic planning system that was resolutely bottom-up. Middle managers were asked to prepare strategic plans for their areas, then at a detailed and lengthy session they presented their thoughts, strategies, requirements and plans to an assembled group of senior managers. These sessions were very one-way—middle managers did all the work leading to the presentations and they did most of the talking. We senior managers sat on the other side of the table and asked occasional questions, which worked to expose weaknesses of logic and inconsistency in the data. But our questions, which were largely nitpicky, didn’t even hint at strategic directions.

  These sessions served their purpose as long as the overarching strategy of the company was simply to produce bigger and better semiconductor memories ahead of the competition. They filled in the details: what technology we needed to develop and how we would develop it, what products we would base on these technologies and so on.

  But as we began to drift into the strategic inflection point that I described in Chapter 5, the woeful inability of this system to deal with big-time change became sorely evident. How could the middle managers who were responsible for memory products deal with the larger question of “Do we even have a chance in the memory business in the first place?” How could the head of the microprocessor organization raise the basic question, “Was it right for Intel to continue to put our best technology resources on the troubled memory business while starving the emerging microprocessor business?” Senior management needed to step in and make some very tough moves, and ultimately, under the duress of lots of red ink, we did. But we also realized then that there must be a better way to formulate strategy.

  What we needed was a balanced interaction between the middle managers, with their deep knowledge but narrow focus, and senior management, whose larger perspective could set a context. The dialectic between these two would often result in searing intellectual debates. But through such debates the shape of the other side of the valley would become clear earlier, making a determined march in its direction more feasible.

  An organization that has a culture that can deal with these two phases—debate (chaos reigns) and a determined march (chaos reined in)—is a powerful, adaptive organization.

  Such an organization has two attributes:

  1. It tolerates and even encourages debate. These debates are vigorous, devoted to exploring issues, indifferent to rank and include individuals of varied backgrounds.

  2. It is capable of making and accepting clear decisions, with the entire organization then supporting the decision.

  Organizations that have these characteristics are far more strategic-inflection-point-ready than others.

  While the description of such a culture is seductively logical, it’s not a very easy environment in which to operate, particularly if you are a newcomer to it and are not familiar with the subtle transition in the swings of the pendulum. An example comes to mind. Some time ago we hired a very competent senior manager from outside Intel as part of the process of bringing computer expertise into our management ranks. He seemed to land on his feet, seemed to enjoy the give-and-take characteristic of our environment and diligently tried to follow the workings of the company as he understood them. Yet he missed the essence of what made it really work.

  At one point he organized a committee and charged it to investigate an issue and come up with a recommendation. It turned out that this manager knew all along what he wanted to do, but instead of giving that direction to the committee, which he could have, he was hoping to engineer a bottom-up decision to the same effect. When the committee came up with the opposite recommendation, he felt cornered. At this late stage, he tried to dictate his solution to people who by now had spent months struggling with an issue and had firmed up their minds. It just couldn’t be done. Coming as it did at this late stage, his dictate seemed utterly arbitrary. The workings of our corporate culture rejected it, and the man had a very hard time understanding where he went wrong.

  The Other Side of the Valley

  Many companies have gone through strategic inflection points. They are alive, competing, even winning. They have survived the challenges of their valleys of death and have emerged stronger than when they went in.

  Hewlett-Packard grew to become a $30 billion company, largely owing to their success in computers. They are second only to IBM in that business.

  Intel became the largest semiconductor producer in the world based on a microprocessor-centered strategy. And Intel coming out of the Pentium processor-flaw crisis is stronger than ever and more in tune with its customers.

  Next is alive and contributing to the computer industry as a software company.

  AT&T and the regional Bell companies thrive, compete and have a market valuation many times what AT&T had before the breakup.

  The ports of Singapore and Seattle are thriving.

  The Warner Brothers movie studio rode the sound wave to become a major media company.

  The other side of the valley of death represents a new industry order that was hard to visualize before the transition. Management did not have a mental map of the new landscape before they encountered it. Getting through the strategic inflection point required enduring a period of confusion, experimentation and chaos, followed by a period of single-minded determination to pursue a new direction toward an initially nebulous goal. It required listening to Cassandras, deliberately fostering debates and constantly articulating the new direction, at first tentatively but more clearly with each repetition. It required casualties and personal transformation; it required accepting the fact that not all would survive and that those who did would not be the same as they had been before.

  Beyond a doubt, going through the valley of death that a strategic inflection point represents is one of the most daunting tasks an organization has to endure. But when “10X” forces are upon us, the choice is taking on these changes or accepting an inevitable decline, which is no choice at all.

  The Internet: Signal or Noise? Threat or Promise?

  “Anything that can

  affect industries

  whose total revenue

  base is many hundreds

  of billions of dollars

  is a big deal.”

  Netscape went public as I was working on this book. I knew of the company, and I thought that they had a lot of promise. But the way the stock skyrocketed on the first day it was available for public purchase and its continuing growth just blew me away. I could find no obvious rational explanation for this incredibly rapid stock appreciation. Something was going on, beyond merely a promising new company being discovered by a growing number of investors.

  Netscape’s business premise was completely intertwined with the evolving Internet. And as the stocks of other Internet-based companies also soared
in Netscape’s wake, it became clear that the investment community’s excitement had as much to do with the Internet as with Netscape.

  The press was not far behind. An avalanche of long feature articles followed, generally creating a dramatized confrontation between software companies that based their activities on the Internet—typified by Netscape, or sometimes Sun—and the established order, as exemplified by Microsoft.

  Something was going on, something was changing.…

  What Is the Internet Anyway?

  For those of you who are not sure what the Internet is but perhaps were afraid to ask, let’s backtrack a bit. Simply put, the Internet is networks of computers connected to each other. If you have a personal computer in California and you’re connected to the Internet, you can exchange data with other computers connected to the Internet in California, in New York, in Germany or in Hong Kong, sort of like what’s shown in the sketch below.

  The work leading to the Internet started in the late sixties with government-initiated—and -funded—connections between various big research computers. The idea was to provide a means of communication that could survive a nuclear explosion that might take down the country’s ordinary telephone infrastructure. Then other computers started joining in. The Internet kept growing and multiplying as people developed more university networks, corporate networks and government networks, and connected them to all the other previously linked networks. They were motivated by the premise that the more computers are connected to each other, the more useful it is to be connected. I like to think of the interlinked networks of all the computers on the Internet as forming a “connection co-op.”