Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 20
RESEARCH RESTRUCTURING AND ASSESSMENT: Can We Apply the Corporate Experience to Government Agencies? 3 Research Restructuring and Assessment at AT&T William F. Brinkman I am going to try a different approach this morning. As I sat and thought about it last night, I started to bring my corporate viewgraphs and then decided that I would not. For a change, I decided not to use viewgraphs but to give a talk that is somewhat more spontaneous and discuss some matters that come to mind. Listening to Jim, I find that the story throughout corporate America today is similar. It just has different local variations. I will start out by talking a little bit about where AT&T is today, because you cannot understand what we are doing in research unless you understand what the corporate institution is doing. OVERVIEW OF AT&T'S ORGANIZATIONAL STRUCTURE As you all know, of course, AT&T was broken up in 1984. It went through about five years of turmoil trying to figure out what it was and how it should be organized. In 1988, we established a new business structure, which has held for the past seven years. AT&T consists of basically four business groups. The first business group is long-distance services; it is, by far, the largest part of AT&T—roughly a $45-billion-a-year business. If you take out the local access charges, it is a $25-billion-a-year business, but nevertheless it is a very large service-oriented business. There is tremendous competition already among AT&T, MCI, and Sprint, but it is very clear that the business will become open to others, particularly the seven regional operating companies. It is just a question of when and how Congress and the Federal Communications Commission (FCC) decide. We believe that there is an inevitability to it. As a corporation we would like to see that there is a relatively level playing field when, in fact, other companies are allowed into the business. That, of course, is where the center of the debate is, what a level playing field should be. The second major business AT&T has is called the network systems business. It is the successor to the old Western Electric, and it supplies communications systems. That business is now international. It was restricted before the Bell system breakup from working in the international arena; its main thrust in the last 11 years has been to become an international player. We have built it up to approximately 33 percent of the international market. It competes with a different set of corporations: Ericsson, Siemens, NEC, Fujitsu, and Alcatel. The last is now the largest telecommunications equipment company in the world. Alcatel has had a very nice arrangement with the French government and has had a lot of cash resources made available to it. Alcatel has bought up a lot of technology by acquiring a number of corporations, and so it has become the biggest supplier of telecommunications equipment in the world. Alcatel is very tough competition today. The third business unit at AT&T is what we call our multimedia business group. This includes the basic handset corded phone business, cordless phones, cellular phones, and the private branch exchange (PBX) systems that are used to control office building communications. We are trying to grow into this part of the corporation—the multimedia area.
OCR for page 21
RESEARCH RESTRUCTURING AND ASSESSMENT: Can We Apply the Corporate Experience to Government Agencies? The fourth business group is what used to be called National Cash Register Company. It is now what we refer to as GIS (Global Information Systems); it was and is a computer company that sells PCs. It sells complete solutions—the same kind of terminology that Jim was using earlier—such as automated teller machines, and is very international. From the start, its business has been about 50 percent domestic and 50 percent foreign. It is a very large company. Finally, we acquired McCaw Cellular Company, which is now part of our service offerings. McCaw is a very large investment in wireless communications. Almost all of the cellular phone business is done at a frequency around 900 megahertz. People, however, believe that digital cellular phones will gradually take over, using a band around 1.9 gigahertz. The FCC just had an auction, and we spent $1.7 billion to buy up bandwidth in the 1.9-gigahertz region in many parts of the country. Its second auction will be coming up in about a year, and we will probably spend another $1.7 billion for additional bandwidth. We are spending another $4 billion to obtain complete ownership of the cellular business of McCaw. When we are finished, we will probably have made a $25-billion investment in wireless communications. The corporation obviously has bet a lot on what wireless communications are going to do in the future. Below these business groups, there are lots of smaller business units. There is a switching business unit, a transmission business unit, a microelectronics business unit, a submarine cable business unit, and so on. The research area works closely with the units at this level. If we look at how we have contributed to those business units, one of the big successes of our research organization in the last 10 years has been in photonics transmission. It is roughly a $2.5-billion-a-year business for AT&T. It has gradually penetrated further and further down from the long-call transmission to the local loop or toward the wire coming into your home. One of the major battles that is shaping up in this industry is access to your home. There are basically two kinds of companies that have access to your home, maybe three if you count the wireless. The first is cable television and the second is your telephone company. In wireless you could add the cellular phone, digital satellite TV, and—perhaps we should not forget—standard broadcast TV. Those are the major services that come into your home today, and there is a coming battle among them. The regional telephone companies are betting enormous amounts of money on getting more bandwidth into your home. I do not know how many of you pay attention to this. Large contracts have been signed among AT&T, Pactel, and a number of the regional companies to essentially introduce broadband services into your house. There are various technical schemes for doing that, but every regional operating company wants to have broadband access and to be able to supply cable television. The cable companies also are trying to figure out how to take their cables and run telephones and data over them. The telephone and cable companies are trying to eat into each other's business. Then there is a third business that is trying to get access to your home—the wireless services, which would scatter some sort of microcell sites around your neighborhood or maybe even on the side of your house, and your initial link to the outside world would be wireless. These are some of the major battles that are looming in telecommunications today. This business is probably close to being at a turning point that is not dissimilar to the introduction of integrated circuits in the electronics business. The upheaval is sufficiently large that —20 or 30 years from now—there will be a totally different set of corporations dominating the business. Of course, we at AT&T are hoping that we are still one of them. But these things are not obvious. Recently there was a major conference on asynchronous transfer mode technology that would create a very flexible packet-based network. Three hundred and fifty corporations showed up at that meeting. So, you can judge how big an issue just one aspect of this technology can be. There are many, many start-ups in the telecommunications world today. About two years ago, Fortune 500 published a list of the hundred fastest-growing companies in the United States. Out of this hundred, 14 of them were in telecommunications. REORGANIZATION OF AT&T RESEARCH There is no question that telecommunications is a very fast-moving field today. We are trying to be part of that field and exert leadership as a research organization. When you listen to these developments, you begin to appreciate that while we might have been doing a broad range of research a long time ago, some of those areas are not as relevant as they used to be. This is a problem that we have consistently faced since divestiture.
OCR for page 22
RESEARCH RESTRUCTURING AND ASSESSMENT: Can We Apply the Corporate Experience to Government Agencies? At the time of divestiture, AT&T's research organization was very science oriented. Three-quarters of it probably was dominated by hardware research of various types. It was more academic than it should have been, and we have made major changes in the last 11 years. We have had to decrease our physical sciences efforts considerably, and we also have had to focus those efforts on the business units. We now have extensive interactions with the business units, and we do that in a number of ways. We have instituted a process of business unit contacts. Out of the 25 subunits that are below the four major business groups I described, each manager in research has a business unit for which he is responsible. He must see to it that the interactions are healthy and that we are doing what that business unit expects. He is expected to meet regularly with the business unit to understand what is going on and to help it. Most of our really good relationships with the business units are so intense that this is an easy job. We have regular quarterly meetings and sometimes monthly meetings. We have subgroups of people that work on specific problems and meet regularly with the business unit. There is no choice about that. If you do not work very closely, you generally are not effective. In the photonics area, of which I have a large portion, we essentially provide the glue among four or five business units. We look at future systems and we try to formulate what those systems are going to look like and then, in fact, sell them to the business units. ASSESSING RESEARCH PERFORMANCE One of the issues that you are dealing with at this workshop and one of the issues that we continually have to deal with is some measurement of what we are doing and how we are doing it. Each year, we put together a list of all our joint projects with all the business units. They do two things with our joint project list. They take the joint projects and look at them, and the business unit forms its own view of how we have served them. That view goes straight to the top of the company and determines our budget. If we do not have good assessments coming up from the business units to the top people, we are in trouble from a budgetary standpoint. Another thing we do is to identify two people, one from the research and from the business unit side, as leaders in every project we have. The person on the business unit side is asked to give us an evaluation of that project. We have a measure by which we have to improve that rating every year. Our salary is determined by whether we improve that number. So, people at my level have a real incentive to see to it that we actually improve our relationships with the business units. We try in research to work the interface with the business unit but, at the same time, keep a balance in the spectrum of research —both the applied research, working with business units, and basic research. If you listen to the story of what our corporation is—that we have just spent $25 billion on McCaw, which is a service corporation, and that our revenues from long distance are $45 billion —you realize that services dominate AT&T. There has been some argument about whether research could be applied to services, but that argument is moot in my opinion. All you have to do today is look at what a chief information officer's job is in a service-oriented company and you realize that there is a tremendous amount of work to be done in database development. For instance, corporations want to have very quick access to all the information they can get on a particular customer. They want to know what is going on at all times with their customers. They want very efficient means of finding out what they are doing themselves. So, there is a tremendous opportunity to come up with ways to improve the efficiency of corporations. There is one story of a corporation that was able to improve the efficiency of its sales force by a factor of three, and to reduce the number of salespeople to that extent, simply by changing its information systems. When you see that, you realize that the new information technology is having a dramatic impact on the efficiency with which organizations can operate and, if you are left behind when somebody else improves by a factor three, you are in trouble. That is the kind of thing we are seeing throughout the industry today. Some people view the competition among AT&T, MCI, and Sprint as a sort of Pepsi-Coke type of thing. If you think about it for a minute, there is not an enormous difference in quality or taste between Coke and Pepsi, but what has maintained the difference has been mainly advertising schemes. So, if you are involved in that kind of battle, if you are not as efficient as your competitor, you have problems because there is very little difference in prices. It is really a matter of making sure that you can meet the competition's price, and research in that area is going to play a very important role. The other thing that we do with respect to joint projects is to look back over a three-year period and come up with some estimates. People really want to assign dollar values to things these days. We have taken our joint
OCR for page 23
RESEARCH RESTRUCTURING AND ASSESSMENT: Can We Apply the Corporate Experience to Government Agencies? projects and tried to come to agreement with the business units on the value of the research we have done over the last few years. This effort has helped us a lot because it has shown how much money really has been made by some of the work we have done. One of our favorites has been the introduction of speech recognition and word spotting in the “zero-plus” calling. I do not know how often you try to go beyond your credit card in zero-plus calling, but you won' t run into an operator today. You reach an automated system that recognizes words. It asks you whether you want to place a collect or a credit card call, and you can answer it in a hundred different ways, but somewhere in the sentence presumably you will use the word “collect” or “credit card” and it will pick that out. That system is something like 98 percent efficient. It is better than the human. So, it has taken over, and a simple estimate of the savings from laying off that many operators is $300 million a year for AT&T. The level of integration on silicon integrated circuits is so high that processing speech is a one-chip issue today for all aspects of speech, thus making the above a practical reality. We are now moving into the video arena where processing a video will be a one-chip solution too, but speech has already been done. I should go back to some of the more traditional aspects of research. We are still interested in and want to maintain a strong scientific base, albeit a smaller one. In order to do that we use fairly conventional means of measuring the value of those people; we have annual reviews. People are funded to start a project, but within a year we begin to look at what has happened. We do not judge proposals. We judge results. If there is one thing I would like to communicate to NSF, it is that mode of operation of judging science. We give people more leeway as they become successful. That is a process we have traditionally used, and everybody who knows Bell Laboratories has heard about our merit review process over the years. It has been a successful means of evaluating science. We also look at prizes. We look at how many people are elected to the National Academy of Sciences and receive other prestigious awards. Another thing we look at is where our people go. If I get a young researcher who stays for 10 years and then goes off to a university, I am not too upset. The reason is that I think I have probably gotten his best 10 years and he will do us a lot of good being out in the university as a friend of the corporation. Recently, I looked back over five years, and it was a very interesting exercise. I restricted myself to physics and looked to see where our physicists had gone. Out of 26 people who had left during that five-year period, 13 of them had gone to the top 10 universities in the country, seven had gone to the top 10 to 20, and the other six were in other universities, but those were the six we had sort of nudged out of the company. I looked at that as a measure of the quality of the people we have had and the quality of their work, and so I found it to be a positive measure. We are using those kinds of measures and I think they are effective. An important issue raised is, How do you change people when you see such a big change in the business environment? We are faced with that problem as much as anyone right now. For instance, we have a tremendous need to build up research in the service areas and in wireless. Fifteen years ago we essentially got out of the wireless business and went into photonics. Today, we cannot get back into the wireless area fast enough, and in fact, the country does not appear to have enough people in that field. I have never seen so much money paid for wireless engineers. This is an area that clearly has a shortage of people, in contrast to many other areas in which we have an overabundance, and so we are trying to address that imbalance in various ways. We are looking at a variety of approaches to retraining, and we are sending some back to school to get master's degrees. I am not sure if our world is changing faster than IBM's, but it certainly is changing very rapidly at this stage, and we want to get out in front on some of these issues. We have been in the forefront of a number of them; certainly, in photonics and in speech recognition, we have done extremely well. We think we have to do better in wireless as a research organization now that we have bought McCaw Cellular. We also think we have to do better in serving the service-oriented business units than we have in the past. GENERAL DISCUSSION DR. MERZ: Bill, I would like to ask a little more about the problem that both you and Jim have referred to, maybe a little obliquely, in the sense of downsizing the basic, unconnected part of your research enterprises. Both IBM and Bell Laboratories had absolutely fantastic research activities 20 years ago. One of the big arguments in those days and today, which you referred to, is the fact that it gives you an opportunity to get very intelligent young people to come into the company, who later discover that this is not how they want to spend the rest of their lives and go on to do the sorts of things that you or your next level down are doing. Right now, to be honest, the word is out,
OCR for page 24
RESEARCH RESTRUCTURING AND ASSESSMENT: Can We Apply the Corporate Experience to Government Agencies? at least to a lot of people, perhaps mistakenly, that those research opportunities are not going to exist, at least not to the degree that they used to. Is that going to hurt you in the long run or what can you do about that? DR. BRINKMAN: That is a subject, which, of course, is a source of some debate, but actually I do not think so. First of all, I would like to say that as we have converted to more applied research, it has been fascinating to watch our researchers because they have been tremendously enthusiastic about what they have done; their enthusiasm is fantastic. They have discovered that the idea of creating real value in the corporation by actually delivering a product, and working on research that the corporation makes a profit on, is an interesting, exciting thing. It is not just the idea of delivering products that is exciting and interesting to them, but that when you get into more applied research, you find that you are dealing with a more complex world. You are dealing with many different variables; it is not like a research project that you have in your lab, over which you have control and about which you write a paper when it is finished. This is a different world. There are many forces operating. People find this aspect of life interesting and find figuring out how to actually deliver a product to be very interesting as well. There is always the “mating game” that goes on at the beginning, but once the conversion occurs, people have jumped in with great enthusiasm. I took a group of people and told them that they would be working on silicon, even though that was not what they had worked on in the past. They were shell-shocked for a month and then decided that this might not be so bad, after all. And over a year or two, they got to be extremely good. We have taken our laser people and said, “making these lasers is great, but they are not the lasers we want. We want the lasers that we can make in production.” We have had to change their mentality. Does that hurt our recruiting? That is a somewhat open question, but I would answer it in the following way. It is arrogant of science to think that the only intelligent people are those who go into science. It seems to me that we should not have that attitude. What I have found over the last several years is that often some of the very best people I get are not people who have Ph.D.s, but people who have electrical engineering degrees and then have worked in industry for a few years. One of my favorite people went to IBM for a few years—not in research, of course—and then went back to school, got a Ph.D., and came to us. He had no inhibition about applied versus basic science, nothing of that sort. He just went to work, and he has had a tremendous impact on the institution in the wireless field, just phenomenal. You know, he doesn't look any less intelligent to me than somebody who worked on electrons in solids when he started out. DR. McGRODDY: Let me make a point about that because one thing we missed before was a complement of that. There are a lot of people that we never saw, that never even interviewed with us or probably Bell Labs, because they did not see in our operations the kind of outlet they felt they needed. We now have the opportunity to work elsewhere in the company, to do start-ups, for instance. We did two start-ups last year directly out of research. This has provided a tremendously broader view of what you can do with your life, and we attract people that we have never seen before. So, I think, in general, we are doing better at recruiting. Part of what is helping, of course, is the fact that the other places that people could go also are less attractive. DR. BROWN: But this is an overly rosy picture, I think, that you are painting. After the fact, after researchers have made the transition, they often find it very exciting, but this mating game is not the fun part. DR. FRIEDMAN: It seems to me that we are beginning here not one workshop, but two. We are talking about two things. One is the change of culture in industry, and we have been talking at the university about it, too. NSF has been looking at that as well. The second, of course, is how we measure performance, which is a different topic. So, it may be that the connection is that we want to take this opportunity of applying measurements to help change the culture. DR. McGRODDY: People do what you measure, certainly in industry. If you put a measurement in place, no matter how silly it is, that is what people will do. DR. FRIEDMAN: But you mentioned, Jim, that since 1989, you started to connect the science to all the application centers and so on. And you mentioned the pleasure people get when they actually go into applications, talk to customers, and find out not only that they are continuing to do research, but also that someone cares about it. DR. BRINKMAN: Yes. I think Jim will agree with me, but I think it is fair to say that industry has struggled with this measurement issue as much as and maybe more than anyone else. It is not an easy issue, and I have concluded that it is very difficult to put any quantitative measure on the science part of your research organization. We have this traditional way of measuring it, which I described in terms of merit review, but to put a number on that is a little tricky. We have tried to make our evaluation more quantitative on the applied side, where we think we know what we are doing and can get some decent measurements, for example, this customer evaluation with the
OCR for page 25
RESEARCH RESTRUCTURING AND ASSESSMENT: Can We Apply the Corporate Experience to Government Agencies? business units. We are trying to understand how that can be improved. It is not easy, and I think the question you are wrestling with today is not an easy issue either because NSF is really supposed to be in basic research. The thing I think that we do, and I would claim NSF does not do, is any systematic evaluation of results, rather than proposals. If you want to come up with a value measure, which is a measurement of how important this science was that resulted from NSF funding, then I think you are going to have to start looking at the results, not just the proposals. DR. McGRODDY: I would like to make one point about measurements that relates to what Erich Bloch said earlier. In industry, particularly, people sometimes try to be quantitative because we know that cost is quantitative. We want a comparison. We have found it very useful to aim high, mainly, not to be satisfied with anything that is not worth 10 times what it costs. So, you can get out of the argument about the precision of measurement, or that you have overestimated or taken double credit. There are many examples where you can get 10, or 100, times the return. DR. LINEBERGER: I want to really emphasize this last comment, this field-of-dreams issue to which I think we must pay very careful attention. If one describes a metric, people will find a way to measure well on it. DR. BLOCH: Just to measure on it, not necessarily well. DR. LINEBERGER: That is true. So, I think that is an important issue that we need to recognize. The other comment is that NSF is also in the education business and I could not help but notice a comment with respect to radio frequency (RF) engineers. You say we do not have enough people to do that research. That is an interesting comment at a time in which people who are in the physical sciences are having difficulty obtaining positions. I wonder if you would like to elaborate at all on that. DR. BRINKMAN: All I can tell you is that I was in a meeting last week with our business units, and I know that we do not have enough people in RF research, although we have tried to hire them. We have had to bid very highly to get them, and all three business units that are involved in wireless hardware were complaining about the fact that they could not get enough people. So, you can say that it is remarkable that we have an overabundance of electrical engineers and physicists and do not have enough RF engineers, but that is a fact. However, I became concerned that somebody would rush off and do what the physicists did. I chaired the NRC physics survey report in the early 1980s. We did a careful analysis of how supply and demand were going to be balanced. We assumed that the physics community would keep producing Ph.D.s at about 950 a year for the next 10 years, and it looked as though there would be a little bit more demand than supply. Well, the universities started producing 1,500 Ph.D.s a year, which means there was a huge surplus, so I want to be careful about this. DR. HACKERMAN: Back in the middle of your talk, you mentioned that you assessed the value of the individual by results. Let's disregard whether you can make a numerical measurement or not. NSF has a different problem. It can assess the results, but it must get people into the system and that is done through proposals. The way you get people into the system is by some sort of a personal examination, individual interviews, et cetera. Some years ago, NSF thought seriously about how it supports individuals and discussed the possibility of doing this simply by assessing the potential that the individual has as a physicist, or whatever, as the MacArthur Foundation does. Another way was providing support to institutions that had demonstrated capabilities. That was rejected because it is an elitist argument and you cannot use public money for elitist purposes. Do you have any suggestions about what NSF can do to improve its intake, rather than worry about an assessment of the results? DR. BRINKMAN: There are two things. One was this idea of funding institutions. Alarms went off the minute you said that because I have dealt so much with Department of Energy (DOE) laboratories, government funding of institutions, and the problems it has led to. It has, first of all, led to essentially a doubling of the bureaucracy because both the laboratories and the DOE have these bureaucracies. DR. HACKERMAN: Well, that has been a problem. There is no question about that. DR. BRINKMAN: So, I would not want NSF to do that. I have sat on committees at NSF where we went through the proposal files and decided whether the refereeing was done well or not. I do not understand why NSF cannot obtain final reports, have some panels look at the final reports, get feedback, and use that feedback in determining future funding. I do not see what is wrong with that. It involves more work. There is one caution I want to make. One of the things we believe in strongly is that we try to promote into scientific research management positions, people who are excellent scientists. I do not believe you can get good judgment without having excellent scientists making the judgment. They know the difference. That is a very important part of it. If you do something, you must be sure that you are using the best people you can to do it.
OCR for page 26
RESEARCH RESTRUCTURING AND ASSESSMENT: Can We Apply the Corporate Experience to Government Agencies? DR. BLOCH: I am singularly unimpressed with your statement that we do not have enough RF engineers and scientists in the country. Anybody who knows what an electron is should be able to become a good RF engineer and scientist in six months. By the way, if they do not, if that transformation cannot take place in a short period of time, that speaks to our education system. What we should have is a measurement of our education system because if you cannot convert people, if they overspecialize—and that is what it sounds like because RF has been around for 50 years or so—then there is something wrong with the education system. We should ask ourselves again what measure we have been using for education, not just for research. DR. FRIEDMAN: It is not easy. DR. FROSCH: It is not what they know or do not know. It is what they believe to be “respectable” to do. They have been miseducated in terms of the respectability of different problems.
Representative terms from entire chapter: