on corporate innovation and what made bell labs successful

innovate or die
innovate or die



i get excited about inventions and innovations, and was digging into how people and insitiutions make these things happen. much has been written on the subject: crossing the chasm, the innovators dilemma, cycles of innovation and discovery, the idea factory, etc. to my dismay, quite a few people believe that we are in a period of great stagnation. the reasons cited for this suspected stagnation are vast and interesting, but are not the topic today. i wanted to uncover the administration and decisions that made places like bell labs great, and to make sure this information isn’t lost in the dramatic, entertaining, and revisionist storytelling that nets book sales.

luckily for us, the vp of research at bell telephone laboratories, ralph brown, gave a talk titled: “vitality of a research instiution and how to maintain it” at the sixth annual conference on the administration of research in 1952. the talk was transcribed and published in the bell telephone systems technical publications journal, monograph 2207 (if you have an original please email me).

i think brown crisply articulates the guard-rails he viewed essential for innovation. the beauty here isn’t just the wisdom, but in how succint and relevant it is, 69 years later. below i’ve pasted the text in it’s entirety:

Vice-President, Research, Bell Telephone Laboratories

IN DEALING WITH THIS SUBJECT I would like first to make clear the scope of my effort and then, avoiding generalities as much as possible, to get down to concrete ideas.

In spite of the fact—or perhaps because of the fact—that Research with a capital R has become a very common word in our language, it is necessary to go a little further than this word alone in order to give a clear understanding of what one is talking about. In my experience the desired understanding is not achieved by prefixing to research such modifying adjectives as basic, fundamental, foundational, applied, product, or any one of the other various qualifying names in common usage. This usually leads merely to long and profitless semantic arguments. For our purposes today let us simply say scientific research to distinguish from literary, economic or other fields, and then content ourselves with saying that we are here concerned with activities which comprise a considerable content of creative work—the uncovering of new knowledge or the achieving of new understandings or, alternatively, inventing new things or making new things work for the first time. Moreover, let us postulate that the activities, while not without concrete practical aims, are nevertheless not narrowly programmed to produce a particular result for a particular delivery date as their major reason for being. In brief, what we are talking about is creative, unscheduled scientific research.

The set of words which I have just uttered is substantially a statement of what we at Bell Telephone Laboratories call research and put into our Research Department, in contradistinction to what we call dedevelopment and design work.

The verb “to maintain” ordinarily carries a connotation of merely supporting, in a state of equilibrium, something already in being. For our purposes I think we must decide that, in a research institution, to maintain vitality implies a dynamic process of continuous growth in which a steady state is achieved only by matching construction against decay. And so perhaps our text would be more informative if it read “Vitality of a Research Institution and How to Build It.”

The vitality of a research organization is only a composite of the spirit of the people in it. It has little to do with buildings or equipment, although indeed these things are important mechanical factors in its existence. In an absolute sense, shelter and tools are indispensable to most research work. But to see thein in their proper relative magnitude one needs to go through a mental exercise somewhat as follows.

Select a research laboratory well known to you and imagine that on a busy morning when all are at work an alarm is suddenly sounded and all the people abandon their tasks and stream out of the building to gather together in a field some distance away. Now let a great explosion wholly destroy the laboratory plant.

Will the instution be killed? Of course it will not. Its functions will be halted and confused for a time, but its vitality, its will to accomplishment, will be challenged to new heights of endeavor and, in borrowed quarters, with makeshift apparatus, it will begin again to carry on.

But now suppose the bomb to fall not on the building but upon the crowd of people gathered in the field. Then indeed the institution would cease to exist. The buildings and plant would be so much inert material. The surrounding populace might be able to supply anew as many live warm human bodies as had previously occupied the structure, but the essence of the institution, its group power in integrated skills of intellect and land, could not thus be recreated.

It seems to me that this mental exercise should throw into relief the fact that the vitality of a large research group is a product of balanced growth; the painstaking selection of those rare individuals who have the gift of uninhibited insight so essential to research in its creative aspects; the equally painstaking shaping of a structure of organization to make use of such people in combination with others of supplementary skills and aptitudes; and, finally, the generation throughout the group of a mutual respect, an esprit de corps and that common will to accomplishment and pride in accomplishment which are in fact its vitality, Building vitality in a research institution is a human problem.

Having thus laid down my statement of the problem, I would like now to go about citing those methods and policies which in my own personal experience I have observed to be effective in achieving solutions.

One of the first things a research institution needs is a technical or scientific objective. Only by having some reasonably well defined goals can researchers make those choices which they face at every tum as to which one of several possible lines to follow. This statement does no violence to the classical concept that true research is without restraint and follows the intellectual curiosity of the researcher. Anyone who has ever done scientific research knows how many interesting paths are continually opening to view in the maze of undiscovered knowledge. Only with a compass can unilateral progress be assured. And without progress of which to be proud there is discouragement and loss of vitality.

The charter does not have to be narrow or constricting to be effective. For example, the improvement of electrical communication as an objective is sufficiently definite to inspire a research laboratory and delineate the fields of science to be cultivated, but it leaves enormous scope and challenge to initiative. One of the vital things for a research organization is to attain a sufficient technical preeminence and fame to attract the admiration and allegiance of desirable members. The charter, by rationalizing and concentrating effort, should help in achieving this end. I believe that those institutions which cover a miscellany of small sponsored projects have a more diflicult problem than do industrial laboratories with better defined lines of interest.

I would not like to be misunderstood on this matter of objective. To this end I would distinguish clearly between the kind of objective which is a broad guiding principle and a thing often called a research project which is an effort to produce a definable result to meet a defined need and, hopefully, at an anticipated time. The Intter thing is an important and necessary feature of many industrial laboratories. I do not decry it. I merely say I am not talking about it.

When a research institution knows its aims and objectives, it automatically knows the fields of scientific progress of which it must become an integral part if it is to be able to grasp and appraise new knowledge and bend it to practical application in the service of man. There is no good way to keep up with the advancing front of scientific knowledge except to be a part of it, to participate actually in the work. This means scientific research within the laboratory itself by employees who are men of a stature and attainment which puts them in equilibrium with the scientific community, men who are known and respected in university and professional society circles and received as equals and peers wherever they find scientists of like interests.

At this point we come to grips with the real problem of vitality. It is simply the problem of how to attract and retain a sufficient number of such men.

The atmosphere of the laboratory and the policy structure on which it rests are important factors in the solution. Therefore let me now enumerate and illustrate some of the components of atmosphere and policy which experience suggests to merit special attention.

First comes that thing commonly called academic freedom. Whether the academic institutions still have an exclusive, or in some cases perhaps even a clear, title to scientific freedom, I do not propose to argue; but it is against whatever they do have that the industrial laboratory must compete. It calls for freedom of publication. Few scientists today seem disposed to protest against the protection of property rights in their work by patent applications, provided these do not unnecessarily delay publication. But trade secrecy is not attractive to them. Interchange of knowledge through publication is the life blood of scientific advance. Any research institution which draws value from the work of others through their publications has in some degree an obligation to return value in kind to the scientific community. The extent to which it does this will weigh heavily in the appraisal which scientists make of it as a place to work.

A great deal could be said about publication policy in detail. This I do not propose to do. But I would like to emphasize that opportunity to publish should be open to all who have something meritorious to tell about and should not be limited to leaders or spokesmen. Only in this atmosphere does the spirit of free scientific enterprise seem to thrive.

Along with free publication as a means of integrating with the scientific community, there go other means of association such as attending technical meetings. visiting other laboratories and entertaining visitors in the home laboratory for discussion of mutually interesting experiments and results. The maintenance of an open door-the freedom of researchers to welcome worthwhile scientific friends and acquaintances and discuss their work, subject to a minimum of necessary restrictions on scope, is a recognized privilege in our most successful laboratories.

Important to an atmosphere of scientific freedom is the organizational structure of an institution. Some research institutions have successfully created the impression that they have no organization charts and that they view such devices with alarm. My own experience leads me to believe that orderliness in relations among people and smoothness in function is facilitated in a large laboratory by a definite charting of responsibilities. The danger comes when the organization chart becomes the master of the organization rather than its servant.

Fostering the development of scientific stature in its technical personnel contributes to the vitality of a laboratory. Scientific stature is something which no organization chart or internal hierarchy of bosses can of themselves produce, since it is the personal contributions of the individual himself as seen by the outside world which determine his scientific stature. The best internal condition is one in which supervisory position and scientific stature are not confused. A scientific leader should feel no need of the trappings of an elevated boss’ job to attain personal satisfaction and financial reward commensurate with his merit. And major supervisors should realize that they have a duty to bring out the best work in others rather than to try to produce all the ideas themselves. There are scientists so gifted and so versatile that they can lead a considerable group both by their own superlative individual contribution and by their influencing and encouraging independent work by their associates. But they are the exception rather than the rule and may well have their own difficulties in the dual role.

There is a widely accepted view that vitality in a research institution requires the genius of a great inspirational leader and that the way to engender such an institution is to identify and catch the right man and then let nature take its course. Just how much the personal qualities, as distinct from the policies of the director or manager, may have to do with the vitality of a research organization I do not propose to attempt to say. Certainly any one who undertakes a director’s job will soon become all too conscious of his own inadequacies. And if he seeks help from the many articles and books which recommend what a director should do and be, he will only become discouraged and convinced that to fill the job is impossible anyway. There is one thing, however, of which I feel surc. Any laboratory which is built around the dominance of its director, however gifted and benevolent he may be, is ill prepared to cope with its future. The best thing a director can do for a research institution is so to shape it that he is not necessary to its vigorous continuity.

What I have said so far has had perhaps more to do with selecting and providing for first class research personnel than it has had with the work which they do. We turn now to the work accomplishment which is the most important thing of all, since it is the end and purpose of the entire establishment. The quality and quantity of the work are a measure of vitality.

In planning and choosing work a research organization needs two kinds of freedom. I have come to refer to these as the two freedoms of industrial research. The first freedom is the freedom to say no to any other group or individual who may come bring ing a problem to which it is proposed that research energy be devoted. It is only through the exercise of independent judgment in selecting work that a line of research endeavor can be consistently, fruitfully and satisfyingly followed. If the research group, including its leadership and management chain of responsibility, is not qualified to make such choices, then the remedy is by changes of personnel rather than by making a research department into a job shop for tough development problems.

Such a basic and important freedom as this cannot in many cases be justified overnight but must be earned over a reasonable period through distinguished performance. Futhermore, its exercise must be tempered by a full recognition of the duty of any research group to put a shoulder to the wheel in common cause during emergencies. Troubles in the factory-a production line held up-ordinarily are things that will be cleared by development or production trouble shooters. But sometimes the trouble is obscure and highly technical, and the whole operation is threatened. Under such circumstances a wise research group will not hesitate to lend its power and value and to throw all its skills into the breach for a limited period.

The existence in a research and development organization of this freedom for research to say no when it deems necessary to protect its own vital functions and energies from dissipation presupposes that develcpment departments are technically strong and well equipped to solve all of their problems which are essentially developmental in character. Research groups should not fear competition from strong development groups but should welcome such strength as the foundation of their own freedom and ability to search in newer fields unhampered by the pressing responsibilities of schedules and commitments. It is from this relationship that I derive the definition of rescarch with which I started this discourse. Develop ment assumes responsibilities for predictable scheduled results. Research guards its freedom to seck new knowledge which development can use. In the detailed technical character of work there may even be considerable areas of temporary overlap.

In freeing himself from the constriction of responsibility for solving development problems, the researcher must concede to the developer the freedom to solve problems in his own way. Without this freedom the development engineer is unable to serve his truc function and may become a mere uninspired routine producer of drawings and specifications. The freedom to select the best way of solving development problems is sometimes called designer’s choice. It implies that the product of research is not accepted perforce but is adopted on merit.

The developer’s freedom to neglect or reject the product of research may, through lack of appreciation, sometimes result in serious failure to utilize available rescarch values. Therefore, the researcher must have a second freedom. This second freedom is the right to carry his ideas experimentally into the applicationa! stage to a point where merit or lack of merit is demonstrated. This point is sometimes difficult to define, but it certainly does not extend into any perfectionizing stage nor does it justify a messianic drive to put over pet personal schemes.

The best protection against the unwise use of the second research freedom is found in the basic responsibility of a research group to remain just that, a research group. The hardest job that researchers, and in particular research supervisors and directors, have is to let go of new things just when they get exciting in an applicational way and turn back to their job of groping for the next good idea. If they fail to exercise this self-control and wisdom and let themselves be drawn into competition with the development engineers, their end as a vital research organism is in sight. Dropping the old to grasp the new is the hardest and most vital decision that research groups are called on to make.

So far I have said little about salaries and labora tory facilities in relation to vitality, not because I fail to recognize that money considerations are a necessary and honorable clement in every job, but because I take it as a premise that to exist and have vital growth a research institution should be in economic equilibrium with its surroundings. That is to say, it should have adequate equipment and be able to pay what the market demands for the kind of services purchased. I do not believe, however, that it can use pay above market levels as a substitute for the kinds of things with which the main body of this paper has been concerned, things which in ordinary industrial parlance would be called good working conditions.

Opportunity for individual growth both in salary and in other satisfactions is essential to a vital organization. A conviction on the part of employees that meritorious performance will be honestly appraised and adequately rewarded is a necessary ingredient of their loyalty. This appraisal, to be fair and convincing, must be based on the individual’s performance and capabilities rather than wholly on the direct value of his results. A system which rewards only those lucky enough to strike an idea which pays off handsomely will not have the cooperative teamwork needed for vitality of the enterprise as a whole.

And now to summarize, I would suggest recognition of the following as important elements in maintaining the vitality of a research institution:

  1. It is primarily a human problem, a problem in human selection, human relations and group spirit;
  2. A well defined but broad technical objective furnishes a rallying point and sharpens deci sions;
  3. The freedom and dignity of the individual in the world of science is a paramount principle;
  4. An orderly organizational structure with room for recognition of a variety of skills is helpful;
  5. A program of successful work which is selfgoverning and which keeps moving dynamically forward into new ground is the purpose of the whole thing;
  6. Just and adequate economic rewards, like good equipment, are a necessary but far from sufficient requirement.
Written on April 13, 2021