Academy of Management, Annual Meeting
Andrew S. Grove
Chairman of the Board, Intel Corporation
San Diego, Calif.
August 9, 1998
Before I start, I have to tell you how I would have answered that question, about the form of management. And I'm not making this up. It was reading Peter Drucker's "Practice of Management" about 30 years after it was published. That kind of proves a couple of points. One is that some of these books are actually read by practitioners, which is reassuring for those of us who write books. And the second one is how long some of the principals that you form active views about last and retain their appropriateness, even as things change. And I suspect several of the things he talked about today will really be seen today as evolutions of his early work.
With that, I would like to tell you a little bit about some thoughts and practices that those of us who work in Intel, in this laboratory of contemporary management, have been involved in over the last year or so.
For those who have not read my last book, I will do a quick review of what Strategic Inflection Points are about. Then I'm going to talk -- borrow something from a colleague of ours, Clayton Christensen from Harvard Business School. And then I'm going to talk about the practical manifestations of both of those ideas to the fundamental change that we deal with these days at Intel.
Let's start with Strategic Inflection Points. They represent, in my description of it, what happens to a business when a major change takes place in its competitive environment. A major change due to introduction of new technologies. A major change due to the introduction of a different regulatory environment. The major change can be simply a change in the customers' values, a change in what customers prefer. Almost always it hits the corporation in such a way that those of us in senior management are among the last ones to notice. I'm paraphrasing the words you used in some of your talk, Peter. But what is common to all of them and what is key is that they require a fundamental change in business strategy, and that's almost a definition of a Strategic Inflection Point. A Strategic Inflection Point is that which causes you to make a fundamental change in business strategy. Nothing less is sufficient.
Being an engineer, forgive me if I like to describe this with a set of curves. An Inflection Point is when a change in the curvature takes place, and depending on the actions you take in responding to this challenge, you will either go on to new heights or head downward in your prosperity as a firm.
The biggest difficulty with Strategic Inflection Points is telling one from the many changes that impinge on you in the business. How do you know if a change is just a garden variety change or qualified to be this monumental, catastrophic change category that we call an Inflection Point? I could never come up with a particularly satisfactory answer to that question. And I'm quite sure that if I could, I wouldn't be talking about them. A set of principles along those lines would be extremely helpful as a competitive weapon because we deal with changes every day. We react to changes every day, but we don't necessarily tear the business up into pieces and put them back together for all of these changes.
Some key warning signs that hint that the change you are dealing with make a Strategic Inflection Point is when it is clear to you that all of a sudden the company or the entity that you worry about has shifted. You have dealt with one particular company or establishment as a competitor all your life and all of a sudden you don't care about them, you care about what somebody else thinks. I have this mental silver bullet test. If you had one bullet, who would you shoot with it? If you change the direction of the gun, that is one of the signals that you may be dealing with something more than an ordinary shift in the competitive landscape.
Likewise, is your key complement a company whose work you rely on to make your company more available? And lastly, when people seem to be talking gibberish around you, people who you worked with for 20 years, who you have a lot of respect for, who normally nobody is talking about and then suddenly everybody seems to talk about them. It's time to sit up and listen and to see what's going on.
Let me give you a few examples of what in hindsight are certainly Strategic Inflection Points. The 1984 modified final judgment that broke off the Bell System has clearly been a Strategic Inflection Point. The telecommunications industry has changed in the next several years in ways that are far more fundamental than anybody could ever imagine. The impact of this change, harking back to Peter's talk, is not what happened in the industry but what happened outside. The telecommunications services became very different from the standpoint of a purchaser, not as a consumer but as a business. You suddenly have to navigate where you have to make different telecommunication vendors' products work together.
In the service businesses, you paid what the Bell System charged you, period. All of this has changed very profoundly. In retailing, we witness in front of our very eyes, all of us as consumers, the impact on the local book sellers by the conglomerations of book sellers -- changes like Barnes & Noble. And if you are one of those local book sellers and Barnes & Noble starts to set up stores in your area -- one, two blocks this way and that way -- you know what a Strategic Inflection Point represents. Your business will never be the same again.
And likewise, in computing, which is closer to my experience, the introduction of PCs was such a Strategic Inflection Point and certainly the manufacturers of mainframes afterward acted exactly like I describe in the book. The senior management was the last to understand what the impact of this change was. Strategic Inflection Points come to us every day.
In telecommunications, the saga continues. In 1996, the Telecommunications Act was introduced by Congress, was voted in by Congress. To almost all of the commentators of the world, it's a bill that signifies nothing. It doesn't change anything. But somehow something must have changed because it is easy to add up something between 150 and 200 billion dollars worth of mergers and acquisition activities that have taken place or that are in the process of taking place in the last two years in the telecommunications industry. The MCI and WorldCom merger. It's an incredible merger. You can't even say it's a merger of equal values in this situation, but a small, more modern company swallowing the bigger company.
The SBC and Pacific Bell, Bell Atlantic and GTE, and most recently, the acquisition of TCI, the cable company, by AT&T. If this isn't a Strategic Inflection Point, then some other natural force has gotten under the skins of the managers of these very large and very slow moving companies to cause them to give up their industry.
In retailing, Barnes & Noble, all of a sudden they have to deal with an Internet-based bookseller upstart that in two years of life has managed to not just eclipse them in sales but eclipse them in sales and market valuation.
Amazon.com, which has a half billion of revenue, has a market value bigger than Barnes & Noble and their next biggest competitor, Borders Inc., put together, and then some. I think the world has changed for Barnes & Noble. They seem to recognize it because they are pushing into online sales -- in fact, they are under-selling themselves on their online efforts, a fairly interesting move. And I don't think it is ever going to be the same for them.
And the last thing: computing. We sit there and scratch our heads but all these digital applications that are sweeping into consumer appliances do represent a Strategic Inflection Point for us, in the same way as the PC for the mainframe. That's not the subject I'm going to talk to you about, but that is another potential example.
There are a variety of tools to use to identify a Strategic Inflection Point. These are empirical observations. One of the empirical observations with companies that are dealing with Strategic Inflection Points, or are in the middle of them, is that there is a growing divergence, a dissonance if you will, between the strategy statements of those companies and the strategy actions. Which of those two adapts to the new realities first? It is strategic actions. Because strategic actions are driven by the competition, driven by the sales force, driven by the sheer necessity of winning business in the marketplace day in and day out. It is a long way before the collection of those strategy actions ends up into the reformulation of what the corporation says it is about.
We have a phrase inside Intel. We are supposed to be a data driven company and the phrase is, "Don't argue with the emotions, argue with the data." And I'm going to say you have to argue with the data at times like this because the data that deal with your business are pertinent to your past, and don't say anything about your future. And the Strategic Inflection Point is a minor activity at present growing in the future and you will never find it in data.
So people who believe that you are dealing with Strategic Inflection Points collide with people who count on winning their argument in these debates based on actual data, and it is like arguing that the change that you are dealing with that is going to transform our future -- the future of the company -- is not real because it wasn't in the past. This is when you see divergence. Our very key phenomenon that has to take place in resolving the Strategic Inflection Point debate between members of management, technologists, and members of the sales force is that very often, because these are not data driven observations, they tend to become emotional arguments. It is extremely important to be able to listen to the people who bring you bad news and who are typically divided.
In the system that I described to you, these people tend to be lower level people. They have to bring you bad news and be Cassandras against the senior management, against the fear of management of repercussions. Unless you deal with this fear, unless you live this fear you will never hear from those helpful Cassandras and you are going to be late in responding to the Strategic Inflection Points.
In dealing with strategic dissonance in dealing with the growing divergent strategies and coping with the phenomenon that every corporation has to deal with, you really go through the stages of grieving. Going from denial, which is the most prominent of all the stages, to various behaviors, to finally acceptance. And once you reach acceptance, action, whether it's sufficient or effective or not, is about to follow.
And it is the very key in all of this to strive toward making these changes as fast as possible because time is your ally. And in turn, your enemy in the world adjusts to the new conditions, and you're going to become a late mover by, for instance, being stuck in the denial stage too long.
I tend to classify the stages of dealing with them in two phases. In the first phase, a lot of experimentation is necessary. So I describe it as letting chaos reign, or experimentation, because you are not likely to start out of the chute and restructure your corporation in the first sign of something.
Rather, you are going let the business units go off and deal with them and adapt to them and watch them like a hawk to see what you are going to learn from them. You are going to watch the dissonance that grows in the company. After letting it grow for a while you start thinking about how you can close it both by changing what you do, and changing what you say you do. And most importantly, you try to picture the new industry -- how the players are going to shape up. What they are going to be looking like at the end of these changes.
As you go through the change, the business goes through major transformation that I can best describe as a Valley of Death. It is not fun to go through those changes. We all say changes are wonderful, we all welcome change. We love change when it happens to somebody else and change is very rarely wonderful when you have to do away with the established practice and established people to adapt -- to tear apart before you can put together the new. So I look at the Valley of Death as an appropriate way to do that.
After you have struggled in the Valley of Death long enough, you have a clear picture of the strategic dissonance and how to close in on it. You are ready for chaos and you're ready to deploy your resources. It's one of those things like empowerment. You want to do it and in reality, when you do it, it is hell.
And one of the disciplines that you need to establish in any organization, with a finite organization, which is any organization, period, you have to balance out the stuff that you want to do, to respond to the new world precisely with the things that you are still doing because the only reason for you to continue to do them would be if the old world continued.
When I say the yeses have to equals the nos, that's what I mean. And everybody is very eager to give you advice how to put yeses to the balance sheet and nobody ever volunteers how to put nos. And really nothing challenges leadership as much -- leadership ability as an organization and the management of the organization rather -- than managing this balance between the yeses and nos.
It is very important to refrain from talking too much about these changes in the early stages. It is important to put the actions in place before you start talking about it because talking prematurely about changes disrupts peoples' lives, existence, employment, and are not truly believed particularly in the positive sense, until a lot of things have begun to be put in practice. But once they are, it is very important for management to come out and speak extremely clearly about what they are going to do and about what these changes are about, and what the company is going to do. At this point, hopefully the other side of the Valley of Death is clear so you can articulate and describe it and you are ready to go. Let's set that aside. Let me talk about the other subject that I would like to call the Christensen Effect. Clayton Christensen has some extremely educational and available drawings in his book titled "The Innovators' Dilemma," from the Harvard Business School Press.
To set the stage up, we have some kind of market performance, something of market values, and that line that you see is the market demand. And you are in the business of providing some kind of capability of product, and your performance is ahead of the market, and behind the market demand. It keeps going up and it's looking very good. At some point, somebody at this point introduces a new approach, a new technology. For instance, Web-based shopping of books. A brand new capability that you look at if you are a Barnes & Noble and say, "God, that is chintzy. That is trivial. You'll never be able to catch me with it. I will be able to provide a service so much better than these people forever."
The Christensen thesis is: It doesn't matter how inferior the new approach is. What matters is, is the new approach is good enough for the market? Does it satisfy the market demand? Long-term success depends on its reaction to phenomenon that are good enough. And just think back to the examples, PCs and mainframes, and the shopping examples, the telecommunications examples, all of these can be thought of along these lines.
But I'm going to go back to Christensen's book and use an example he developed there which is the onset of the minimills and the integrated steel mills. The integrated steel mills have supplied subsegments of the steel market -- structure steel, various bars and rods all the way to the lowest end of that, which is rebars. And they happily did this and dealt with our union work force and did whatever they were doing. And then someplace in the late ‘70s, the minimills started to come in providing small amounts of fairly low-quality steel, but it was quality that was good enough for rebars.
And in short order, the steel company said that's not exactly a business that is so important to us anyway. It was a low margin business. Rough and dirty business. Hell with it. Let them have it. And there the story would end, except the minimills improved their quality, and got into the other types of bars, and improved the quality some more, and got into structure steel, and improved their quality some more -- to the point where at last the product is good enough for the last refuge of the steel segments, which is the steel plate industry. Christensen gave a seminar at Intel, and I was looking at his slide as I was talking about the industry, and I had this creeping feeling in my gut that what he was really talking about was the computer industry and the most important characteristic -- which is the incident introduction of low-cost computers that are somehow different than the low-cost computers used to be.
And as we sit here and fully dismiss the characteristics and the markets, the low cost, we are sounding just like the proverbial integrated steel companies we're talking about when they were looking at the loss of the rebar market. This is what I would like to call the Segment Zero Phenomenon. You'll see that in a minute. But I think it is a good way to look at it.
You'll see why I call it Segment Zero. Let me talk a little bit about the low-cost PC, the history of the PC. To explain to you why it was so easy for us to dismiss it. There have been PCs forever and they were crummy. And they were crummy because the paradigm in which they operated, they used to last years, and don't take that literally. It could have been two years old. They have a low performance, limited features, applications that cost people to buy, and they all failed. So we never included this category of PCs in our own market. This is the Intel triangle market model.
Segment Two has a higher performance and higher and so on and so forth, and, of course, as you went up higher segments, PCs were more expensive. They appeal to a smaller and smaller population. The target got narrower and narrower.
What the data told us, it was first of all that the market was like this forever and ever and didn't change. By the way, there was that little category of those low-cost PCs. We are going to call them Segment Zero. They fall off our market value. That's what it stands for. It falls off your market value. It is small. It is obsolete. It is bottom fishing. It is our rebar. And generally, we were okay with this data. We're talking about a little over a year ago. In ‘97, the segments are basically healthy even though there are some changes on them.
These are generally Intel data. This is what we were looking at. First of all, it has to do only with the U.S. retail. Desktop, 10 percent of the total worldwide volume, so it's not that important. But it's very visible that that's the stuff that's advertising in your Sunday newspaper. You sip your coffee and you read about the 10 percent of the marketplace and you think the whole world is like that, etc. etc. It's easy to dismiss. But not that easy to dismiss because the newspaper rubs your eyes into it, and also the financial analyst will find it easier to go down to the local Comp U.S.A than to go to Asia to look at the market.
We keep reporting about this segment as if it was much larger. What this chart does is show a progression going back to 1994. Segment Zero is the dark green, Segment One is the next set of bars. Segment Two is the next set of bars and so on and so forth. And you see Segment Zero seems to be growing. It was down to 3, 4, or 5 percent in U.S. retail. But it's now up to 20 percent.
It's a cycle. Look back to where it was in 1994. After that, memory prices went up. All the prices got pushed up and you couldn't afford to build the low cost PC because of the component costs. The same thing is happening now. There's an excess of memory, an excess of this and that. We are going back to historical patterns.
This seems to be confirmed by looking at average selling prices. There's a little bit of a drop. You are not going to sound the trumpets for a $100 difference. However, maybe this is what was going to happen, you never know.
One interesting thing is, it is very difficult to control experiments in the marketplace. This is terribly frustrating for a technology company. We'll never know what would have happened at the time of the Asia crisis. It didn't come in the middle third quarter of the same year, ‘97. And that triggered a bunch of price wars. Why did the Asian price crisis trigger a price war? It started a cycle by taking a big chunk of the growth out of the PC industry. Our combined growth had come out of very rapid growth in Asian consumption. As that disappeared, the sales started to drop.
There was a market share war. Market share wars, first of all, led to stuffing of inventory on shelves which in turn exacerbated the price war. All of a sudden the pricing spiral became a whole lot worse.
Correspondingly, the U.S. average selling price in the U.S. retail channel dropped $100 or so. Prices started to accelerate downward. When we put all these things together, we had to face a very disturbing possibility that this market model that had served us for decades, may be changing. In the combination of technical changes I'll talk about in a moment, and the Asian phenomenon fueling those technical changes, we may have changed the market distribution shape.
So at this point, when we started fighting, we heard Clayton Christensen's rebar story. So rebars became kind of a slogan to those who listened to seminars. We are going to start fighting the rebar area.
Well, the first way we started to fight is by price. We started charging for plates at rebar prices and that is not necessarily a lasting strategy. But it worked. The bad news is it accelerated what we called -- another internal Intel jargon -- the waterfall. What is the waterfall? After this, you will be able to fake your way into any Intel strategy session with the proper linking of these phrases.
Waterfall represents how we used to operate. Introduce a microprocessor in the high end of the market, then a year later we would come down in the midrange. Then a year or two later, come down to the very lowest end of the phenomenon. It trickled down. And then, of course, as it trickled down, it created a terrible hole for us to introduce a new generation. The cycle was like waterfall. It replenished every season.
The transition from the top to the bottom took about three years. That has been the way of this industry for a long time. Remember, signs of a Strategic Inflection Point are changes that fundamentally change the industry and cause you to do your business differently. Take a look at this. In the new waterfall, you put the data in and the change necessitated this whole transition from top to bottom. That used to take three years, and now it happens in 12 to 18 months.
Under those circumstances, Segment Zero products, which never worked before because they were always yesterday's product, all of a sudden became good enough because they were not really yesterday's product. The stuff that we introduced the first of ‘97 actually waterfalled down to the Segment Zero segment by the end of 1997. It's very difficult to call the machine that you marketed worldwide as the best thing since sliced cheese in January an obsolete product by the time December roles around.
So we needed to change. Segment Zero was in fact a Strategic Inflection Point. We had to make some changes as to how we operated. Let me talk about that in the last section of the talk, and let me tell you some of the actions that we took in dealing with this phenomenon. And again, I remind you, this is a work in process. It's not over. It's probably never going to be over, but as of now, this is what we have brought to bear on this problem.
We've stopped arguing with the data. We segmented our products forever. I'm going to talk about this at length in a moment. We adopted a branch structure. We adopted a whole different attitude to our product, including cost improvements, and we had to change the managerial attitudes and the whole product line one at a time.
Let's talk about the data. I mentioned this earlier. Data tells you more about the past than about the future. But data is a wonderful, objective, thing to hide behind when some parts of your organization tell you that you have to do something tremendously difficult. You can always hide behind the data. Until you stop dealing with this, you will not get anywhere. We had to go back to some of the fundamental belief systems. And one of the things we have to believe in is the notion that prices never go up, because we helped create it. Once they have gone down, we are very likely to have to deal with that without being able to count on recovery. The recovery that you noticed in the dip happened a few years earlier, but it happened under an industry shortage, supply shortage, and balance mis-shift. Anybody who counts on that kind of a thing as a strategy element is living a fairly unreasonable and unexamined strategy life.
So we had to realize that the market is, in fact, segmenting and the old way of treating it as a uniform market that we can deal with and handle through the waterfall won't work anymore. And when we started to step back and look at the historical changes in our market, we realized that long before this date somebody kicked our shin out with this Segment Zero phenomenon. We were just trying to paint all this with the same brush and service it with a simple product strategy.
The PC used to be one thing. The PC already has stopped being one thing. There are PCs that were used for servicing data that had different characteristics and different manufacturers and different customers and different value sets that customers wanted. The business PC had different value sets from the consumer PC, which had different technical and different customers from the mobile PCs. We looked forward to looking at our own product developments, not completed plans but thought plans. We saw that a segmentation was going to cut each of those and cut them into an even larger number, at least twice the number of segments.
Most importantly, we notice that there's a category in each of these products -- conceptual product road maps. For example, there's the Basic PC under the business segment, sometimes called a lean client, and the Basic PC for the consumer segment. Segment Zero categories clearly have to be dealt with as separate segments. And if it is a separate segment, it has a separate value set. We need to take those into account and have our product -- our own product, which is the microprocessor which goes into the PC -- contemplate it right from the beginning. Or we need to design it for the target segment, rather than reproduce it.
So once we internalized this, in one quarter, we took 650 engineers and put them to work on this Segment Zero microprocessor road map. These are products designed specifically from our best ideas on how to achieve what buyers of Segment Zero PCs wanted. We have a certain number of transistors that we can put into our design. As technology gets better, we can put more transistors on a chip. Ordinarily, we use those to improve the performance of chips. We continue to do that for the performance segment. What we will use is extra components to reduce the cost of the total PC, not just the cost of the chip, for the manufacturer by integrating additional functions on that same chip.
Technically, this is as valid a use of the extra transistors as the others. Simply, we have never done it because we ever never considered Segment Zero part of our market .
Given that we have segmented our product road map, we segmented our brand road map. We have been very, very pleased with ourselves for having created this. In a three-year period of time, the Intel® Pentium® processor brand has become one of the strongest product brands in the world. Very few brands are created as profusely and as uniquely and as quickly as the Intel Pentium processor. We're very much attached to it.
As part of our segmentation, we broke our reliance on the Intel Pentium processor brand for all of the segments and introduced two different brands. We introduced the Intel Pentium® II Xeon processor brand for the high-end server market, which also requires different products.
We pulled off a miracle by introducing the Intel Pentium processor. Are we counting on doubling that miracle? Do we have enough money to do that? Are we willing to advertise this? Is there going to be interaction between those brands? Are we going to confuse the markets? There's more questions that you can ask to which the answer could be, "Let's just play it safe and hang on to the brand that we established." Well, we didn't. And it remains to be seen how successful the multitude of brands is going to be. But that's what we did.
Lastly, we have realized long ago that in our industry, you must price the market and cost must follow price. This is a major difference between the computing industry and the telecommunications industry, but we think of price and market relationships completely in the opposite fashion. And we needed to reaffirm our belief in this and start going with cost reduction efforts to make this actually possible, because the market has spoken. We have, by legitimizing Segment Zero, accepted the market. It has spoken on the subject, and we now have to speak in our factories. This was our actual and planned costs -- weighted average cost. By the time we put solid plans together, this is what cost phenomenon brought on our design.
Lastly, managerial attitude had to change from denial to acceptance when talking about Segment Zero -- it's something here to stay. The very word implies Zero is something that doesn't warrant attention by putting it on a road map. We learned to live with it. We have a dedicated team to service it and we pursue it with the cost fanaticism that our industry has normally employed.
To give you a roadmap, frankly, the trade newspapers have basically covered everything I told you here. Let me quote a phrase from "Electronic Buyers' News." ["Having secured a beachhead in the mainstream PC arena, Intel Corp. is adopting an increasingly sophisticated strategy of segmentation as it renews its push into the server and low-cost PC markets." -- Aug. 3, 1998] This is kind of good for me to read because it comes from a very cynical, jaundiced newspaper that typically targets purchasers and managers. So the people who write for the computing system seem to think we are really doing this. I certainly think that we are doing it, but I'm a top manager. I'm subject to self-deception, so I prefer to believe "Electronic Buyers' News."
Now, the last thought I want to leave you with is the Segment Zero phenomenon is not a computing phenomenon. It is not an Intel PC phenomenon. It is much broader. If you start looking at it, and probably a lot of you in this room have dealt with the industry of each of these and taught classes on them, in terms of looking for examples of PC segments of the computing industry, they were treated as such. No mainframe company put it on the road map. And none of them adopted the kind of phenomenon. None of them have become sustaining leaders in the new industry. Low-cost personal copiers appeared from a Japanese company to the U.S. industry as a Segment Zero of the industry. Another example is the compact segment of the car industry in the ‘70s. And as we look around, the Internet is a major conveyance of the Segment Zero phenomenon in all kinds of different areas. Stock trade is a Segment Zero of the stock brokerage business. It is at a much lower cost with poorer qualities. The advice you get and the service you get relative to having your personal stockbroker is not the same.
The question to ask is: Is it good enough? Is it good enough for a large segment? Is the segment for whom it is good enough growing? And if the answer to that is yes, is it going to cause a Strategic Inflection Point for the brokerage industry? Or the Internet and the data that it has used to date -- is it good enough for now? The answer is no. I can almost imagine that providers of this service say, this is a joke. Like CB radios. This is like rebars. And like minimills, this technology is going to get better, and it is going to be good enough for an increasing population. And if so, it is going to bring a Strategic Inflection Point for the ordinary telephone company.
Direct selling online and Pentium II Xeon processor .com style leaves a lot to be desired. Or going down to your neighborhood store and browsing again in the book stores -- is it good enough? If it is good enough, a growing segment of people will get used to purchasing books that way. Cars can't be far behind. Can other things be far behind? And if not, will this be a Strategic Inflection Point for the retail industry? And the problem with this is Segment Zero creates the opportunity for new market leaders to emerge. They have no sheet steel or high-cost, high-performance PCs to depend on. They are putting all of their efforts in that, and they pose a huge challenge to the incumbent and require dealing with this phenomenon of Strategic Inflection Point.
I described to you some of the things that we have done. Only time will tell whether they are sufficient or whether we too will be subject to the same degradation as those ahead of us who haven't.
Last thought, before you jump out and call every change a Segment Zero phenomenon. Not everything that's low cost is a Segment Zero. Not everything that's low cost is good enough to pay for or has a chance and opportunity of getting better on a secular basis. And not everything is likely to be embraced by a good portion of the buying population.
So much as was the case with Strategic Inflection Points. I wish I could give another couple of visuals, slides here, showing you what some of the criteria are. But likely with other forms of Strategic Inflection Point, that's where art comes in and the science goes by the wayside.
Thank you very much.
* Other names and brands may be claimed as the property of others.